notebooks/2-ml-models.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Libraries\n",
    "import pandas as pd, numpy as np, seaborn as sn, matplotlib.pyplot as plt, ipywidgets as widgets\n",
    "from utils import *\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn import tree\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.svm import SVC\n",
    "from sklearn import metrics, preprocessing"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Credit Risk Analysis - Machine Learning models <br>\n",
    "\n",
    "The objective of this analysis is to understand which are the factors that are more effecting the probability of having an high credit risk of a given company, using initially only the data of the explanatory variables of previous year to predict the credit risk of the consequent one (true values vs forecasts). <br>\n",
    "Then it can be added new explanatory variables 'delayed' e.g. leverage of two years, three years, (...) before. <br>\n",
    "After understood some general behavior (through the previous notebook), this specific notebook has the objective to build different models to predict the credit risk level (low/high) in an year for each company, and also compare each presented model (in order to both interpret them and see their accuracy).\n",
    "\n",
    "**Financial Data Science (UniPV) - Prof. Paolo Giudici** <br>\n",
    "*Computer Engineering (Data Science) - A.Y. 2022/23* <br>\n",
    "*Francesco Amato, 507767*"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import the dataset & data manipulations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Company name</th>\n",
       "      <th>Turnover.2020</th>\n",
       "      <th>Turnover.2019</th>\n",
       "      <th>Turnover.2018</th>\n",
       "      <th>Turnover.2017</th>\n",
       "      <th>Turnover.2016</th>\n",
       "      <th>Turnover.2015</th>\n",
       "      <th>EBIT.2020</th>\n",
       "      <th>EBIT.2019</th>\n",
       "      <th>EBIT.2018</th>\n",
       "      <th>...</th>\n",
       "      <th>Materials</th>\n",
       "      <th>Media &amp; Entertainment</th>\n",
       "      <th>Pharmaceuticalcs biotechnology and life sci</th>\n",
       "      <th>Real Estate</th>\n",
       "      <th>Retailing</th>\n",
       "      <th>Software and Services</th>\n",
       "      <th>Technology Hardware and Equipment</th>\n",
       "      <th>Telecommunication Services</th>\n",
       "      <th>Transportation</th>\n",
       "      <th>Utilities</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>LENDLEASE S.R.L.</td>\n",
       "      <td>29458</td>\n",
       "      <td>16716</td>\n",
       "      <td>9612</td>\n",
       "      <td>8097</td>\n",
       "      <td>7941.0</td>\n",
       "      <td>5600.0</td>\n",
       "      <td>-1556.0</td>\n",
       "      <td>-4540.0</td>\n",
       "      <td>623.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PRICEWATERHOUSECOOPERS BUSINESS SERVICES SRL (...</td>\n",
       "      <td>16731</td>\n",
       "      <td>16403</td>\n",
       "      <td>16843</td>\n",
       "      <td>12241</td>\n",
       "      <td>9252.0</td>\n",
       "      <td>9515.0</td>\n",
       "      <td>1838.0</td>\n",
       "      <td>841.0</td>\n",
       "      <td>2738.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>EVISO S.P.A.</td>\n",
       "      <td>48568</td>\n",
       "      <td>43039</td>\n",
       "      <td>34302</td>\n",
       "      <td>25791</td>\n",
       "      <td>19760.0</td>\n",
       "      <td>6941.0</td>\n",
       "      <td>1661.0</td>\n",
       "      <td>1464.0</td>\n",
       "      <td>976.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CASA SERVICE MACHINE</td>\n",
       "      <td>47999</td>\n",
       "      <td>43484</td>\n",
       "      <td>43043</td>\n",
       "      <td>41682</td>\n",
       "      <td>51267.0</td>\n",
       "      <td>52584.0</td>\n",
       "      <td>416.0</td>\n",
       "      <td>255.0</td>\n",
       "      <td>-855.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>PANFERTIL SPA</td>\n",
       "      <td>45948</td>\n",
       "      <td>47336</td>\n",
       "      <td>45626</td>\n",
       "      <td>48222</td>\n",
       "      <td>57074.0</td>\n",
       "      <td>62263.0</td>\n",
       "      <td>44.0</td>\n",
       "      <td>713.0</td>\n",
       "      <td>-672.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121003</th>\n",
       "      <td>ASTOR VILLAGE S.R.L.</td>\n",
       "      <td>3161</td>\n",
       "      <td>4635</td>\n",
       "      <td>4742</td>\n",
       "      <td>4499</td>\n",
       "      <td>4277.0</td>\n",
       "      <td>3650.0</td>\n",
       "      <td>985.0</td>\n",
       "      <td>1818.0</td>\n",
       "      <td>1790.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121004</th>\n",
       "      <td>ODONE &amp; SLOA S.R.L.</td>\n",
       "      <td>3161</td>\n",
       "      <td>2562</td>\n",
       "      <td>2559</td>\n",
       "      <td>2334</td>\n",
       "      <td>3692.0</td>\n",
       "      <td>2537.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>101.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121005</th>\n",
       "      <td>GARRIDO MURO SOCIEDAD LIMITADA</td>\n",
       "      <td>3161</td>\n",
       "      <td>3146</td>\n",
       "      <td>2989</td>\n",
       "      <td>3101</td>\n",
       "      <td>2746.0</td>\n",
       "      <td>3154.0</td>\n",
       "      <td>260.0</td>\n",
       "      <td>13.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121006</th>\n",
       "      <td>CENTRO INGROSSO JOLLY S.R.L.</td>\n",
       "      <td>3161</td>\n",
       "      <td>2519</td>\n",
       "      <td>2290</td>\n",
       "      <td>2244</td>\n",
       "      <td>1761.0</td>\n",
       "      <td>1821.0</td>\n",
       "      <td>74.0</td>\n",
       "      <td>48.0</td>\n",
       "      <td>60.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121007</th>\n",
       "      <td>SALONES COMATEL SL</td>\n",
       "      <td>3161</td>\n",
       "      <td>4514</td>\n",
       "      <td>4435</td>\n",
       "      <td>4231</td>\n",
       "      <td>3908.0</td>\n",
       "      <td>2051.0</td>\n",
       "      <td>194.0</td>\n",
       "      <td>733.0</td>\n",
       "      <td>830.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>121008 rows × 79 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             Company name  Turnover.2020  \\\n",
       "0                                        LENDLEASE S.R.L.          29458   \n",
       "1       PRICEWATERHOUSECOOPERS BUSINESS SERVICES SRL (...          16731   \n",
       "2                                            EVISO S.P.A.          48568   \n",
       "3                                    CASA SERVICE MACHINE          47999   \n",
       "4                                           PANFERTIL SPA          45948   \n",
       "...                                                   ...            ...   \n",
       "121003                               ASTOR VILLAGE S.R.L.           3161   \n",
       "121004                                ODONE & SLOA S.R.L.           3161   \n",
       "121005                     GARRIDO MURO SOCIEDAD LIMITADA           3161   \n",
       "121006                       CENTRO INGROSSO JOLLY S.R.L.           3161   \n",
       "121007                                 SALONES COMATEL SL           3161   \n",
       "\n",
       "        Turnover.2019  Turnover.2018  Turnover.2017  Turnover.2016  \\\n",
       "0               16716           9612           8097         7941.0   \n",
       "1               16403          16843          12241         9252.0   \n",
       "2               43039          34302          25791        19760.0   \n",
       "3               43484          43043          41682        51267.0   \n",
       "4               47336          45626          48222        57074.0   \n",
       "...               ...            ...            ...            ...   \n",
       "121003           4635           4742           4499         4277.0   \n",
       "121004           2562           2559           2334         3692.0   \n",
       "121005           3146           2989           3101         2746.0   \n",
       "121006           2519           2290           2244         1761.0   \n",
       "121007           4514           4435           4231         3908.0   \n",
       "\n",
       "        Turnover.2015  EBIT.2020  EBIT.2019  EBIT.2018  ...  Materials  \\\n",
       "0              5600.0    -1556.0    -4540.0      623.0  ...          0   \n",
       "1              9515.0     1838.0      841.0     2738.0  ...          0   \n",
       "2              6941.0     1661.0     1464.0      976.0  ...          0   \n",
       "3             52584.0      416.0      255.0     -855.0  ...          0   \n",
       "4             62263.0       44.0      713.0     -672.0  ...          0   \n",
       "...               ...        ...        ...        ...  ...        ...   \n",
       "121003         3650.0      985.0     1818.0     1790.0  ...          0   \n",
       "121004         2537.0       60.0      101.0       27.0  ...          0   \n",
       "121005         3154.0      260.0       13.0       48.0  ...          0   \n",
       "121006         1821.0       74.0       48.0       60.0  ...          0   \n",
       "121007         2051.0      194.0      733.0      830.0  ...          0   \n",
       "\n",
       "        Media & Entertainment  Pharmaceuticalcs biotechnology and life sci  \\\n",
       "0                           0                                            0   \n",
       "1                           0                                            0   \n",
       "2                           0                                            0   \n",
       "3                           0                                            0   \n",
       "4                           0                                            0   \n",
       "...                       ...                                          ...   \n",
       "121003                      0                                            0   \n",
       "121004                      0                                            0   \n",
       "121005                      0                                            0   \n",
       "121006                      0                                            0   \n",
       "121007                      0                                            0   \n",
       "\n",
       "        Real Estate  Retailing  Software and Services  \\\n",
       "0                 0          0                      0   \n",
       "1                 0          0                      0   \n",
       "2                 0          0                      0   \n",
       "3                 0          1                      0   \n",
       "4                 0          1                      0   \n",
       "...             ...        ...                    ...   \n",
       "121003            0          0                      0   \n",
       "121004            0          0                      0   \n",
       "121005            0          0                      0   \n",
       "121006            0          1                      0   \n",
       "121007            0          0                      0   \n",
       "\n",
       "        Technology Hardware and Equipment  Telecommunication Services  \\\n",
       "0                                       0                           0   \n",
       "1                                       0                           0   \n",
       "2                                       0                           0   \n",
       "3                                       0                           0   \n",
       "4                                       0                           0   \n",
       "...                                   ...                         ...   \n",
       "121003                                  0                           0   \n",
       "121004                                  0                           0   \n",
       "121005                                  0                           0   \n",
       "121006                                  0                           0   \n",
       "121007                                  0                           0   \n",
       "\n",
       "        Transportation Utilities  \n",
       "0                    0         0  \n",
       "1                    0         0  \n",
       "2                    0         1  \n",
       "3                    0         0  \n",
       "4                    0         0  \n",
       "...                ...       ...  \n",
       "121003               0         0  \n",
       "121004               0         0  \n",
       "121005               0         0  \n",
       "121006               0         0  \n",
       "121007               0         0  \n",
       "\n",
       "[121008 rows x 79 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Load the datatset\n",
    "companies_df = pd.read_csv('../datasets/one-hot-credit-risk.tar.gz', compression='gzip')\n",
    "companies_df"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Balancing the companies (use same number for high and low level credit risk)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Fix the unbalanced case\n",
    "high_risk_df = companies_df[companies_df['MScore.2019.int'] == 1]\n",
    "low_risk_df = companies_df[companies_df['MScore.2019.int'] == 0].sample(n=len(high_risk_df), random_state=0)\n",
    "restricted_df = pd.concat([low_risk_df, high_risk_df])\n",
    "restricted_df.sort_index(inplace=True)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Consider past year company data to predict next year credit risk level"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Select the year to use to predict the consequent one credit risk level\n",
    "\n",
    "**Beware of the year choice**, because of the already mentioned Covid (possible) effect."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c33fe400649044e784ed0aecc74291a1",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "IntSlider(value=2018, description='Year:', max=2019, min=2015)"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "year_widget = widgets.IntSlider(value=2018, min=2015, max=2019, description='Year:', disabled=False, continuous_update=True, orientation='horizontal', readout=True, readout_format='d')\n",
    "year_widget"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### First objective: feature selection\n",
    "\n",
    "Since we want to just understand which are the features that are more helpful to improve the predictive accuracy, and the other performance metrics, we will start considering a very easy model (**Logistic Regression**). <br>\n",
    "Once found the most impactful features we will see whether considering other more complex models (mantaining fixed the features) the performance metrics will change signficantly (or not)."
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Easiest model: credit risk level of previous year, to predict the one of next year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2018.int</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120994</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120995</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120998</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121003</th>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121004</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 1 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        MScore.2018.int\n",
       "0                     1\n",
       "4                     1\n",
       "7                     0\n",
       "11                    0\n",
       "16                    0\n",
       "...                 ...\n",
       "120994                0\n",
       "120995                1\n",
       "120998                0\n",
       "121003                0\n",
       "121004                1\n",
       "\n",
       "[24320 rows x 1 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = restricted_df[['MScore.'+str(year_widget.value)+'.int']]\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 78.7%\n",
      " - Test accuracy: 79.4%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.72      0.96      0.82      2432\n",
      "           1       0.94      0.63      0.75      2432\n",
      "\n",
      "    accuracy                           0.79      4864\n",
      "   macro avg       0.83      0.79      0.79      4864\n",
      "weighted avg       0.83      0.79      0.79      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Consideration:** The overall obtained performances can be considered good, but what if we add and consider other features?<br>"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Add another feature: credit risk level + turnover of previous year, to predict the one of next year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2018.int</th>\n",
       "      <th>Turnover.2018</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.158884</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.912978</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.856485</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.701328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.641379</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24315</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.032581</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24316</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.048264</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24317</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.026216</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24318</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.056912</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24319</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.011202</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       MScore.2018.int  Turnover.2018\n",
       "0                  1.0       0.158884\n",
       "1                  1.0       0.912978\n",
       "2                  0.0       0.856485\n",
       "3                  0.0       0.701328\n",
       "4                  0.0       0.641379\n",
       "...                ...            ...\n",
       "24315              0.0       0.032581\n",
       "24316              1.0       0.048264\n",
       "24317              0.0       0.026216\n",
       "24318              0.0       0.056912\n",
       "24319              1.0       0.011202\n",
       "\n",
       "[24320 rows x 2 columns]"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = restricted_df[['MScore.'+str(year_widget.value)+'.int', 'Turnover.'+str(year_widget.value)]]\n",
    "X = pd.DataFrame(preprocessing.MinMaxScaler().fit_transform(X), columns=X.columns) # MinMax Scaling\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 78.7%\n",
      " - Test accuracy: 79.4%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.72      0.96      0.82      2432\n",
      "           1       0.94      0.63      0.75      2432\n",
      "\n",
      "    accuracy                           0.79      4864\n",
      "   macro avg       0.83      0.79      0.79      4864\n",
      "weighted avg       0.83      0.79      0.79      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Add another feature: credit risk score + EBIT of previous year, to predict the one of next year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2018.int</th>\n",
       "      <th>EBIT.2018</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.777574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.772238</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.791195</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.776359</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.790593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24315</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.775291</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24316</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.775386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24317</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.776107</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24318</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.782384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24319</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.775118</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       MScore.2018.int  EBIT.2018\n",
       "0                  1.0   0.777574\n",
       "1                  1.0   0.772238\n",
       "2                  0.0   0.791195\n",
       "3                  0.0   0.776359\n",
       "4                  0.0   0.790593\n",
       "...                ...        ...\n",
       "24315              0.0   0.775291\n",
       "24316              1.0   0.775386\n",
       "24317              0.0   0.776107\n",
       "24318              0.0   0.782384\n",
       "24319              1.0   0.775118\n",
       "\n",
       "[24320 rows x 2 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = restricted_df[['MScore.'+str(year_widget.value)+'.int', 'EBIT.'+str(year_widget.value)]]\n",
    "X = pd.DataFrame(preprocessing.MinMaxScaler().fit_transform(X), columns=X.columns) # MinMax Scaling\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 78.7%\n",
      " - Test accuracy: 79.4%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.72      0.96      0.82      2432\n",
      "           1       0.94      0.63      0.75      2432\n",
      "\n",
      "    accuracy                           0.79      4864\n",
      "   macro avg       0.83      0.79      0.79      4864\n",
      "weighted avg       0.83      0.79      0.79      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Add another feature: credit risk level + PLTax of previous year, to predict the one of next year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2018.int</th>\n",
       "      <th>PLTax.2018</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.809337</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.806730</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.821109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.809904</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.820282</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24315</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.809366</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24316</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.809299</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24317</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.810130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24318</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.814631</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24319</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.809249</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       MScore.2018.int  PLTax.2018\n",
       "0                  1.0    0.809337\n",
       "1                  1.0    0.806730\n",
       "2                  0.0    0.821109\n",
       "3                  0.0    0.809904\n",
       "4                  0.0    0.820282\n",
       "...                ...         ...\n",
       "24315              0.0    0.809366\n",
       "24316              1.0    0.809299\n",
       "24317              0.0    0.810130\n",
       "24318              0.0    0.814631\n",
       "24319              1.0    0.809249\n",
       "\n",
       "[24320 rows x 2 columns]"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = restricted_df[['MScore.'+str(year_widget.value)+'.int', 'PLTax.'+str(year_widget.value)]]\n",
    "X = pd.DataFrame(preprocessing.MinMaxScaler().fit_transform(X), columns=X.columns) # MinMax Scaling\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 78.7%\n",
      " - Test accuracy: 79.4%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.72      0.96      0.82      2432\n",
      "           1       0.94      0.63      0.75      2432\n",
      "\n",
      "    accuracy                           0.79      4864\n",
      "   macro avg       0.83      0.79      0.79      4864\n",
      "weighted avg       0.83      0.79      0.79      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Add another feature: credit risk level of previous year + country, to predict the one of next year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2018.int</th>\n",
       "      <th>France</th>\n",
       "      <th>Germany</th>\n",
       "      <th>Italy</th>\n",
       "      <th>Spain</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120994</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120995</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120998</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121003</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121004</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        MScore.2018.int  France  Germany  Italy  Spain\n",
       "0                     1       0        0      1      0\n",
       "4                     1       0        0      1      0\n",
       "7                     0       0        0      1      0\n",
       "11                    0       1        0      0      0\n",
       "16                    0       0        0      1      0\n",
       "...                 ...     ...      ...    ...    ...\n",
       "120994                0       0        0      1      0\n",
       "120995                1       0        0      1      0\n",
       "120998                0       1        0      0      0\n",
       "121003                0       0        0      1      0\n",
       "121004                1       0        0      1      0\n",
       "\n",
       "[24320 rows x 5 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = pd.concat([restricted_df[['MScore.'+str(year_widget.value)+'.int']], restricted_df.loc[:, 'France':'Spain']], axis=1)\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 78.7%\n",
      " - Test accuracy: 79.4%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.72      0.96      0.82      2432\n",
      "           1       0.94      0.63      0.75      2432\n",
      "\n",
      "    accuracy                           0.79      4864\n",
      "   macro avg       0.83      0.79      0.79      4864\n",
      "weighted avg       0.83      0.79      0.79      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Add another feature: credit risk level of previous year + general sector, to predict the one of next year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2018.int</th>\n",
       "      <th>Automobiles and Components</th>\n",
       "      <th>Capital Goods</th>\n",
       "      <th>Commercial and professional services</th>\n",
       "      <th>Consumer Durables and Apparel</th>\n",
       "      <th>Consumer Services</th>\n",
       "      <th>Diversified Financials</th>\n",
       "      <th>Energy</th>\n",
       "      <th>Food Beverage and Tobacco</th>\n",
       "      <th>Food and Staples Retailing</th>\n",
       "      <th>...</th>\n",
       "      <th>Materials</th>\n",
       "      <th>Media &amp; Entertainment</th>\n",
       "      <th>Pharmaceuticalcs biotechnology and life sci</th>\n",
       "      <th>Real Estate</th>\n",
       "      <th>Retailing</th>\n",
       "      <th>Software and Services</th>\n",
       "      <th>Technology Hardware and Equipment</th>\n",
       "      <th>Telecommunication Services</th>\n",
       "      <th>Transportation</th>\n",
       "      <th>Utilities</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120994</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120995</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120998</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121003</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121004</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 22 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        MScore.2018.int  Automobiles and Components  Capital Goods  \\\n",
       "0                     1                           0              1   \n",
       "4                     1                           0              0   \n",
       "7                     0                           0              1   \n",
       "11                    0                           0              0   \n",
       "16                    0                           0              1   \n",
       "...                 ...                         ...            ...   \n",
       "120994                0                           0              1   \n",
       "120995                1                           0              0   \n",
       "120998                0                           0              0   \n",
       "121003                0                           0              0   \n",
       "121004                1                           0              1   \n",
       "\n",
       "        Commercial and professional services  Consumer Durables and Apparel  \\\n",
       "0                                          0                              0   \n",
       "4                                          0                              0   \n",
       "7                                          0                              0   \n",
       "11                                         0                              1   \n",
       "16                                         0                              0   \n",
       "...                                      ...                            ...   \n",
       "120994                                     0                              0   \n",
       "120995                                     0                              0   \n",
       "120998                                     0                              0   \n",
       "121003                                     0                              0   \n",
       "121004                                     0                              0   \n",
       "\n",
       "        Consumer Services  Diversified Financials  Energy  \\\n",
       "0                       0                       0       0   \n",
       "4                       0                       0       0   \n",
       "7                       0                       0       0   \n",
       "11                      0                       0       0   \n",
       "16                      0                       0       0   \n",
       "...                   ...                     ...     ...   \n",
       "120994                  0                       0       0   \n",
       "120995                  0                       0       0   \n",
       "120998                  1                       0       0   \n",
       "121003                  1                       0       0   \n",
       "121004                  0                       0       0   \n",
       "\n",
       "        Food Beverage and Tobacco  Food and Staples Retailing  ...  Materials  \\\n",
       "0                               0                           0  ...          0   \n",
       "4                               0                           0  ...          0   \n",
       "7                               0                           0  ...          0   \n",
       "11                              0                           0  ...          0   \n",
       "16                              0                           0  ...          0   \n",
       "...                           ...                         ...  ...        ...   \n",
       "120994                          0                           0  ...          0   \n",
       "120995                          0                           1  ...          0   \n",
       "120998                          0                           0  ...          0   \n",
       "121003                          0                           0  ...          0   \n",
       "121004                          0                           0  ...          0   \n",
       "\n",
       "        Media & Entertainment  Pharmaceuticalcs biotechnology and life sci  \\\n",
       "0                           0                                            0   \n",
       "4                           0                                            0   \n",
       "7                           0                                            0   \n",
       "11                          0                                            0   \n",
       "16                          0                                            0   \n",
       "...                       ...                                          ...   \n",
       "120994                      0                                            0   \n",
       "120995                      0                                            0   \n",
       "120998                      0                                            0   \n",
       "121003                      0                                            0   \n",
       "121004                      0                                            0   \n",
       "\n",
       "        Real Estate  Retailing  Software and Services  \\\n",
       "0                 0          0                      0   \n",
       "4                 0          1                      0   \n",
       "7                 0          0                      0   \n",
       "11                0          0                      0   \n",
       "16                0          0                      0   \n",
       "...             ...        ...                    ...   \n",
       "120994            0          0                      0   \n",
       "120995            0          0                      0   \n",
       "120998            0          0                      0   \n",
       "121003            0          0                      0   \n",
       "121004            0          0                      0   \n",
       "\n",
       "        Technology Hardware and Equipment  Telecommunication Services  \\\n",
       "0                                       0                           0   \n",
       "4                                       0                           0   \n",
       "7                                       0                           0   \n",
       "11                                      0                           0   \n",
       "16                                      0                           0   \n",
       "...                                   ...                         ...   \n",
       "120994                                  0                           0   \n",
       "120995                                  0                           0   \n",
       "120998                                  0                           0   \n",
       "121003                                  0                           0   \n",
       "121004                                  0                           0   \n",
       "\n",
       "        Transportation  Utilities  \n",
       "0                    0          0  \n",
       "4                    0          0  \n",
       "7                    0          0  \n",
       "11                   0          0  \n",
       "16                   0          0  \n",
       "...                ...        ...  \n",
       "120994               0          0  \n",
       "120995               0          0  \n",
       "120998               0          0  \n",
       "121003               0          0  \n",
       "121004               0          0  \n",
       "\n",
       "[24320 rows x 22 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = pd.concat([restricted_df[['MScore.'+str(year_widget.value)+'.int']], restricted_df.loc[:, 'Automobiles and Components':'Utilities']], axis=1)\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 78.7%\n",
      " - Test accuracy: 79.4%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.72      0.96      0.82      2432\n",
      "           1       0.94      0.63      0.75      2432\n",
      "\n",
      "    accuracy                           0.79      4864\n",
      "   macro avg       0.83      0.79      0.79      4864\n",
      "weighted avg       0.83      0.79      0.79      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Add another feature: credit risk level + Leverage of previous year, to predict the one of next year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2018.int</th>\n",
       "      <th>Leverage.2018</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.086342</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.086212</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.086208</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.086215</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.086213</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24315</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.086229</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24316</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.086246</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24317</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.086209</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24318</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.086208</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24319</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.086268</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       MScore.2018.int  Leverage.2018\n",
       "0                  1.0       0.086342\n",
       "1                  1.0       0.086212\n",
       "2                  0.0       0.086208\n",
       "3                  0.0       0.086215\n",
       "4                  0.0       0.086213\n",
       "...                ...            ...\n",
       "24315              0.0       0.086229\n",
       "24316              1.0       0.086246\n",
       "24317              0.0       0.086209\n",
       "24318              0.0       0.086208\n",
       "24319              1.0       0.086268\n",
       "\n",
       "[24320 rows x 2 columns]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = restricted_df[['MScore.'+str(year_widget.value)+'.int', 'Leverage.'+str(year_widget.value)]]\n",
    "X = pd.DataFrame(preprocessing.MinMaxScaler().fit_transform(X), columns=X.columns) # MinMax Scaling\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 78.7%\n",
      " - Test accuracy: 79.4%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.72      0.96      0.82      2432\n",
      "           1       0.94      0.63      0.75      2432\n",
      "\n",
      "    accuracy                           0.79      4864\n",
      "   macro avg       0.83      0.79      0.79      4864\n",
      "weighted avg       0.83      0.79      0.79      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Add another feature: credit risk level + ROE of previous year, to predict the one of next year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2018.int</th>\n",
       "      <th>ROE.2018</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.152948</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.152937</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.152946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.152954</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.152966</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24315</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.152947</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24316</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.152946</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24317</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.152980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24318</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.152950</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24319</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.152942</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       MScore.2018.int  ROE.2018\n",
       "0                  1.0  0.152948\n",
       "1                  1.0  0.152937\n",
       "2                  0.0  0.152946\n",
       "3                  0.0  0.152954\n",
       "4                  0.0  0.152966\n",
       "...                ...       ...\n",
       "24315              0.0  0.152947\n",
       "24316              1.0  0.152946\n",
       "24317              0.0  0.152980\n",
       "24318              0.0  0.152950\n",
       "24319              1.0  0.152942\n",
       "\n",
       "[24320 rows x 2 columns]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = restricted_df[['MScore.'+str(year_widget.value)+'.int', 'ROE.'+str(year_widget.value)]]\n",
    "X = pd.DataFrame(preprocessing.MinMaxScaler().fit_transform(X), columns=X.columns) # MinMax Scaling\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 78.7%\n",
      " - Test accuracy: 79.4%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.72      0.96      0.82      2432\n",
      "           1       0.94      0.63      0.75      2432\n",
      "\n",
      "    accuracy                           0.79      4864\n",
      "   macro avg       0.83      0.79      0.79      4864\n",
      "weighted avg       0.83      0.79      0.79      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Add another feature: credit risk level + TAsset of previous year, to predict the one of next year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2018.int</th>\n",
       "      <th>TAsset.2018</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.007275</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.017136</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.031393</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.005507</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.014624</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24315</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.001933</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24316</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.001715</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24317</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000322</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24318</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.006792</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24319</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.001048</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       MScore.2018.int  TAsset.2018\n",
       "0                  1.0     0.007275\n",
       "1                  1.0     0.017136\n",
       "2                  0.0     0.031393\n",
       "3                  0.0     0.005507\n",
       "4                  0.0     0.014624\n",
       "...                ...          ...\n",
       "24315              0.0     0.001933\n",
       "24316              1.0     0.001715\n",
       "24317              0.0     0.000322\n",
       "24318              0.0     0.006792\n",
       "24319              1.0     0.001048\n",
       "\n",
       "[24320 rows x 2 columns]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = restricted_df[['MScore.'+str(year_widget.value)+'.int', 'TAsset.'+str(year_widget.value)]]\n",
    "X = pd.DataFrame(preprocessing.MinMaxScaler().fit_transform(X), columns=X.columns) # MinMax Scaling\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 78.7%\n",
      " - Test accuracy: 79.4%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.72      0.96      0.82      2432\n",
      "           1       0.94      0.63      0.75      2432\n",
      "\n",
      "    accuracy                           0.79      4864\n",
      "   macro avg       0.83      0.79      0.79      4864\n",
      "weighted avg       0.83      0.79      0.79      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Combine all features together: credit risk level + Turnover + EBIT + PLTax + Sector + Country + Leverage + ROE + TAsset of previous year, to predict the one of the next year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2018.int</th>\n",
       "      <th>Turnover.2018</th>\n",
       "      <th>EBIT.2018</th>\n",
       "      <th>PLTax.2018</th>\n",
       "      <th>Leverage.2018</th>\n",
       "      <th>ROE.2018</th>\n",
       "      <th>TAsset.2018</th>\n",
       "      <th>Automobiles and Components</th>\n",
       "      <th>Capital Goods</th>\n",
       "      <th>Commercial and professional services</th>\n",
       "      <th>...</th>\n",
       "      <th>Retailing</th>\n",
       "      <th>Software and Services</th>\n",
       "      <th>Technology Hardware and Equipment</th>\n",
       "      <th>Telecommunication Services</th>\n",
       "      <th>Transportation</th>\n",
       "      <th>Utilities</th>\n",
       "      <th>France</th>\n",
       "      <th>Germany</th>\n",
       "      <th>Italy</th>\n",
       "      <th>Spain</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.158884</td>\n",
       "      <td>0.777574</td>\n",
       "      <td>0.809337</td>\n",
       "      <td>0.086342</td>\n",
       "      <td>0.152948</td>\n",
       "      <td>0.007275</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.912978</td>\n",
       "      <td>0.772238</td>\n",
       "      <td>0.806730</td>\n",
       "      <td>0.086212</td>\n",
       "      <td>0.152937</td>\n",
       "      <td>0.017136</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.856485</td>\n",
       "      <td>0.791195</td>\n",
       "      <td>0.821109</td>\n",
       "      <td>0.086208</td>\n",
       "      <td>0.152946</td>\n",
       "      <td>0.031393</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.701328</td>\n",
       "      <td>0.776359</td>\n",
       "      <td>0.809904</td>\n",
       "      <td>0.086215</td>\n",
       "      <td>0.152954</td>\n",
       "      <td>0.005507</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.641379</td>\n",
       "      <td>0.790593</td>\n",
       "      <td>0.820282</td>\n",
       "      <td>0.086213</td>\n",
       "      <td>0.152966</td>\n",
       "      <td>0.014624</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24315</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.032581</td>\n",
       "      <td>0.775291</td>\n",
       "      <td>0.809366</td>\n",
       "      <td>0.086229</td>\n",
       "      <td>0.152947</td>\n",
       "      <td>0.001933</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24316</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.048264</td>\n",
       "      <td>0.775386</td>\n",
       "      <td>0.809299</td>\n",
       "      <td>0.086246</td>\n",
       "      <td>0.152946</td>\n",
       "      <td>0.001715</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24317</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.026216</td>\n",
       "      <td>0.776107</td>\n",
       "      <td>0.810130</td>\n",
       "      <td>0.086209</td>\n",
       "      <td>0.152980</td>\n",
       "      <td>0.000322</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24318</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.056912</td>\n",
       "      <td>0.782384</td>\n",
       "      <td>0.814631</td>\n",
       "      <td>0.086208</td>\n",
       "      <td>0.152950</td>\n",
       "      <td>0.006792</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24319</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.011202</td>\n",
       "      <td>0.775118</td>\n",
       "      <td>0.809249</td>\n",
       "      <td>0.086268</td>\n",
       "      <td>0.152942</td>\n",
       "      <td>0.001048</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 32 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       MScore.2018.int  Turnover.2018  EBIT.2018  PLTax.2018  Leverage.2018  \\\n",
       "0                  1.0       0.158884   0.777574    0.809337       0.086342   \n",
       "1                  1.0       0.912978   0.772238    0.806730       0.086212   \n",
       "2                  0.0       0.856485   0.791195    0.821109       0.086208   \n",
       "3                  0.0       0.701328   0.776359    0.809904       0.086215   \n",
       "4                  0.0       0.641379   0.790593    0.820282       0.086213   \n",
       "...                ...            ...        ...         ...            ...   \n",
       "24315              0.0       0.032581   0.775291    0.809366       0.086229   \n",
       "24316              1.0       0.048264   0.775386    0.809299       0.086246   \n",
       "24317              0.0       0.026216   0.776107    0.810130       0.086209   \n",
       "24318              0.0       0.056912   0.782384    0.814631       0.086208   \n",
       "24319              1.0       0.011202   0.775118    0.809249       0.086268   \n",
       "\n",
       "       ROE.2018  TAsset.2018  Automobiles and Components  Capital Goods  \\\n",
       "0      0.152948     0.007275                         0.0            1.0   \n",
       "1      0.152937     0.017136                         0.0            0.0   \n",
       "2      0.152946     0.031393                         0.0            1.0   \n",
       "3      0.152954     0.005507                         0.0            0.0   \n",
       "4      0.152966     0.014624                         0.0            1.0   \n",
       "...         ...          ...                         ...            ...   \n",
       "24315  0.152947     0.001933                         0.0            1.0   \n",
       "24316  0.152946     0.001715                         0.0            0.0   \n",
       "24317  0.152980     0.000322                         0.0            0.0   \n",
       "24318  0.152950     0.006792                         0.0            0.0   \n",
       "24319  0.152942     0.001048                         0.0            1.0   \n",
       "\n",
       "       Commercial and professional services  ...  Retailing  \\\n",
       "0                                       0.0  ...        0.0   \n",
       "1                                       0.0  ...        1.0   \n",
       "2                                       0.0  ...        0.0   \n",
       "3                                       0.0  ...        0.0   \n",
       "4                                       0.0  ...        0.0   \n",
       "...                                     ...  ...        ...   \n",
       "24315                                   0.0  ...        0.0   \n",
       "24316                                   0.0  ...        0.0   \n",
       "24317                                   0.0  ...        0.0   \n",
       "24318                                   0.0  ...        0.0   \n",
       "24319                                   0.0  ...        0.0   \n",
       "\n",
       "       Software and Services  Technology Hardware and Equipment  \\\n",
       "0                        0.0                                0.0   \n",
       "1                        0.0                                0.0   \n",
       "2                        0.0                                0.0   \n",
       "3                        0.0                                0.0   \n",
       "4                        0.0                                0.0   \n",
       "...                      ...                                ...   \n",
       "24315                    0.0                                0.0   \n",
       "24316                    0.0                                0.0   \n",
       "24317                    0.0                                0.0   \n",
       "24318                    0.0                                0.0   \n",
       "24319                    0.0                                0.0   \n",
       "\n",
       "       Telecommunication Services  Transportation  Utilities  France  Germany  \\\n",
       "0                             0.0             0.0        0.0     0.0      0.0   \n",
       "1                             0.0             0.0        0.0     0.0      0.0   \n",
       "2                             0.0             0.0        0.0     0.0      0.0   \n",
       "3                             0.0             0.0        0.0     1.0      0.0   \n",
       "4                             0.0             0.0        0.0     0.0      0.0   \n",
       "...                           ...             ...        ...     ...      ...   \n",
       "24315                         0.0             0.0        0.0     0.0      0.0   \n",
       "24316                         0.0             0.0        0.0     0.0      0.0   \n",
       "24317                         0.0             0.0        0.0     1.0      0.0   \n",
       "24318                         0.0             0.0        0.0     0.0      0.0   \n",
       "24319                         0.0             0.0        0.0     0.0      0.0   \n",
       "\n",
       "       Italy  Spain  \n",
       "0        1.0    0.0  \n",
       "1        1.0    0.0  \n",
       "2        1.0    0.0  \n",
       "3        0.0    0.0  \n",
       "4        1.0    0.0  \n",
       "...      ...    ...  \n",
       "24315    1.0    0.0  \n",
       "24316    1.0    0.0  \n",
       "24317    0.0    0.0  \n",
       "24318    1.0    0.0  \n",
       "24319    1.0    0.0  \n",
       "\n",
       "[24320 rows x 32 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = restricted_df[['MScore.'+str(year_widget.value)+'.int', 'Turnover.'+str(year_widget.value), 'EBIT.'+str(year_widget.value), 'PLTax.'+str(year_widget.value), 'Leverage.'+str(year_widget.value), 'ROE.'+str(year_widget.value), 'TAsset.'+str(year_widget.value)]]\n",
    "X = pd.concat([X, restricted_df.loc[:, 'Automobiles and Components':'Utilities']], axis=1)\n",
    "X = pd.concat([X, restricted_df.loc[:, 'France':'Spain']], axis=1)\n",
    "X = pd.DataFrame(preprocessing.MinMaxScaler().fit_transform(X), columns=X.columns) # MinMax Scaling\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 78.7%\n",
      " - Test accuracy: 79.4%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.72      0.96      0.82      2432\n",
      "           1       0.94      0.63      0.75      2432\n",
      "\n",
      "    accuracy                           0.79      4864\n",
      "   macro avg       0.83      0.79      0.79      4864\n",
      "weighted avg       0.83      0.79      0.79      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Consider also older than one year company data to predict consequent year credit risk level"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Models\n",
    "\n",
    "#### All past years credit risk levels to predict the one of the next year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2015.int</th>\n",
       "      <th>MScore.2016.int</th>\n",
       "      <th>MScore.2017.int</th>\n",
       "      <th>MScore.2018.int</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120994</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120995</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120998</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121003</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121004</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        MScore.2015.int  MScore.2016.int  MScore.2017.int  MScore.2018.int\n",
       "0                     1                0                1                1\n",
       "4                     0                0                1                1\n",
       "7                     0                0                0                0\n",
       "11                    0                0                0                0\n",
       "16                    0                0                0                0\n",
       "...                 ...              ...              ...              ...\n",
       "120994                1                1                0                0\n",
       "120995                1                1                1                1\n",
       "120998                0                0                0                0\n",
       "121003                0                0                0                0\n",
       "121004                1                1                1                1\n",
       "\n",
       "[24320 rows x 4 columns]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for i in range(0, year_widget.value - 2014):\n",
    "    if i == 0: X = restricted_df[['MScore.2015.int']]\n",
    "    else: X = pd.concat([X, restricted_df[['MScore.'+str(2015 + i)+'.int']]], axis=1)\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 81.5%\n",
      " - Test accuracy: 82.4%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.77      0.91      0.84      2432\n",
      "           1       0.89      0.73      0.81      2432\n",
      "\n",
      "    accuracy                           0.82      4864\n",
      "   macro avg       0.83      0.82      0.82      4864\n",
      "weighted avg       0.83      0.82      0.82      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### All features together: credit risk level + Turnover + EBIT + PLTax + Sector + Country + Leverage + ROE + TAsset of also past years, to predict the one of the next year"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2015.int</th>\n",
       "      <th>Turnover.2015</th>\n",
       "      <th>EBIT.2015</th>\n",
       "      <th>PLTax.2015</th>\n",
       "      <th>Leverage.2015</th>\n",
       "      <th>ROE.2015</th>\n",
       "      <th>TAsset.2015</th>\n",
       "      <th>MScore.2016.int</th>\n",
       "      <th>Turnover.2016</th>\n",
       "      <th>EBIT.2016</th>\n",
       "      <th>...</th>\n",
       "      <th>Retailing</th>\n",
       "      <th>Software and Services</th>\n",
       "      <th>Technology Hardware and Equipment</th>\n",
       "      <th>Telecommunication Services</th>\n",
       "      <th>Transportation</th>\n",
       "      <th>Utilities</th>\n",
       "      <th>France</th>\n",
       "      <th>Germany</th>\n",
       "      <th>Italy</th>\n",
       "      <th>Spain</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.026555</td>\n",
       "      <td>0.567529</td>\n",
       "      <td>0.577861</td>\n",
       "      <td>0.001091</td>\n",
       "      <td>0.878438</td>\n",
       "      <td>0.005735</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.029056</td>\n",
       "      <td>0.302037</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.295250</td>\n",
       "      <td>0.575646</td>\n",
       "      <td>0.577941</td>\n",
       "      <td>0.001087</td>\n",
       "      <td>0.878846</td>\n",
       "      <td>0.020328</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.208834</td>\n",
       "      <td>0.299928</td>\n",
       "      <td>...</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.111328</td>\n",
       "      <td>0.579888</td>\n",
       "      <td>0.586102</td>\n",
       "      <td>0.001087</td>\n",
       "      <td>0.878858</td>\n",
       "      <td>0.031902</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.162903</td>\n",
       "      <td>0.310055</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.127450</td>\n",
       "      <td>0.572621</td>\n",
       "      <td>0.581604</td>\n",
       "      <td>0.001087</td>\n",
       "      <td>0.878888</td>\n",
       "      <td>0.003837</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.113188</td>\n",
       "      <td>0.300696</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.088964</td>\n",
       "      <td>0.577920</td>\n",
       "      <td>0.584856</td>\n",
       "      <td>0.001087</td>\n",
       "      <td>0.878875</td>\n",
       "      <td>0.010428</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.072324</td>\n",
       "      <td>0.304515</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24315</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.020372</td>\n",
       "      <td>0.571538</td>\n",
       "      <td>0.580967</td>\n",
       "      <td>0.001087</td>\n",
       "      <td>0.878830</td>\n",
       "      <td>0.002160</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.015203</td>\n",
       "      <td>0.299668</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24316</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.020191</td>\n",
       "      <td>0.572749</td>\n",
       "      <td>0.581634</td>\n",
       "      <td>0.001088</td>\n",
       "      <td>0.878856</td>\n",
       "      <td>0.001881</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.015759</td>\n",
       "      <td>0.299949</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24317</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.014297</td>\n",
       "      <td>0.573039</td>\n",
       "      <td>0.582162</td>\n",
       "      <td>0.001087</td>\n",
       "      <td>0.878879</td>\n",
       "      <td>0.000568</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.011248</td>\n",
       "      <td>0.300147</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24318</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.017308</td>\n",
       "      <td>0.575030</td>\n",
       "      <td>0.583288</td>\n",
       "      <td>0.001087</td>\n",
       "      <td>0.878859</td>\n",
       "      <td>0.005753</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.015650</td>\n",
       "      <td>0.302757</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24319</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.012030</td>\n",
       "      <td>0.572640</td>\n",
       "      <td>0.581717</td>\n",
       "      <td>0.001088</td>\n",
       "      <td>0.878842</td>\n",
       "      <td>0.001551</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.013509</td>\n",
       "      <td>0.299692</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 53 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       MScore.2015.int  Turnover.2015  EBIT.2015  PLTax.2015  Leverage.2015  \\\n",
       "0                  1.0       0.026555   0.567529    0.577861       0.001091   \n",
       "1                  0.0       0.295250   0.575646    0.577941       0.001087   \n",
       "2                  0.0       0.111328   0.579888    0.586102       0.001087   \n",
       "3                  0.0       0.127450   0.572621    0.581604       0.001087   \n",
       "4                  0.0       0.088964   0.577920    0.584856       0.001087   \n",
       "...                ...            ...        ...         ...            ...   \n",
       "24315              1.0       0.020372   0.571538    0.580967       0.001087   \n",
       "24316              1.0       0.020191   0.572749    0.581634       0.001088   \n",
       "24317              0.0       0.014297   0.573039    0.582162       0.001087   \n",
       "24318              0.0       0.017308   0.575030    0.583288       0.001087   \n",
       "24319              1.0       0.012030   0.572640    0.581717       0.001088   \n",
       "\n",
       "       ROE.2015  TAsset.2015  MScore.2016.int  Turnover.2016  EBIT.2016  ...  \\\n",
       "0      0.878438     0.005735              0.0       0.029056   0.302037  ...   \n",
       "1      0.878846     0.020328              0.0       0.208834   0.299928  ...   \n",
       "2      0.878858     0.031902              0.0       0.162903   0.310055  ...   \n",
       "3      0.878888     0.003837              0.0       0.113188   0.300696  ...   \n",
       "4      0.878875     0.010428              0.0       0.072324   0.304515  ...   \n",
       "...         ...          ...              ...            ...        ...  ...   \n",
       "24315  0.878830     0.002160              1.0       0.015203   0.299668  ...   \n",
       "24316  0.878856     0.001881              1.0       0.015759   0.299949  ...   \n",
       "24317  0.878879     0.000568              0.0       0.011248   0.300147  ...   \n",
       "24318  0.878859     0.005753              0.0       0.015650   0.302757  ...   \n",
       "24319  0.878842     0.001551              1.0       0.013509   0.299692  ...   \n",
       "\n",
       "       Retailing  Software and Services  Technology Hardware and Equipment  \\\n",
       "0            0.0                    0.0                                0.0   \n",
       "1            1.0                    0.0                                0.0   \n",
       "2            0.0                    0.0                                0.0   \n",
       "3            0.0                    0.0                                0.0   \n",
       "4            0.0                    0.0                                0.0   \n",
       "...          ...                    ...                                ...   \n",
       "24315        0.0                    0.0                                0.0   \n",
       "24316        0.0                    0.0                                0.0   \n",
       "24317        0.0                    0.0                                0.0   \n",
       "24318        0.0                    0.0                                0.0   \n",
       "24319        0.0                    0.0                                0.0   \n",
       "\n",
       "       Telecommunication Services  Transportation  Utilities  France  Germany  \\\n",
       "0                             0.0             0.0        0.0     0.0      0.0   \n",
       "1                             0.0             0.0        0.0     0.0      0.0   \n",
       "2                             0.0             0.0        0.0     0.0      0.0   \n",
       "3                             0.0             0.0        0.0     1.0      0.0   \n",
       "4                             0.0             0.0        0.0     0.0      0.0   \n",
       "...                           ...             ...        ...     ...      ...   \n",
       "24315                         0.0             0.0        0.0     0.0      0.0   \n",
       "24316                         0.0             0.0        0.0     0.0      0.0   \n",
       "24317                         0.0             0.0        0.0     1.0      0.0   \n",
       "24318                         0.0             0.0        0.0     0.0      0.0   \n",
       "24319                         0.0             0.0        0.0     0.0      0.0   \n",
       "\n",
       "       Italy  Spain  \n",
       "0        1.0    0.0  \n",
       "1        1.0    0.0  \n",
       "2        1.0    0.0  \n",
       "3        0.0    0.0  \n",
       "4        1.0    0.0  \n",
       "...      ...    ...  \n",
       "24315    1.0    0.0  \n",
       "24316    1.0    0.0  \n",
       "24317    0.0    0.0  \n",
       "24318    1.0    0.0  \n",
       "24319    1.0    0.0  \n",
       "\n",
       "[24320 rows x 53 columns]"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for i in range(0, year_widget.value - 2014):\n",
    "    if i == 0: X = restricted_df[['MScore.2015.int', 'Turnover.2015', 'EBIT.2015', 'PLTax.2015', 'Leverage.2015', 'ROE.2015', 'TAsset.2015']]\n",
    "    else: X = pd.concat([X, restricted_df[['MScore.'+str(2015 + i)+'.int', 'Turnover.'+str(2015 + i), 'EBIT.'+str(2015 + i), 'PLTax.'+str(2015 + i), 'Leverage.'+str(2015 + i), 'ROE.'+str(2015 + i), 'TAsset.'+str(2015 + i)]]], axis=1)\n",
    "X = pd.concat([X, restricted_df.loc[:, 'Automobiles and Components':'Utilities']], axis=1)\n",
    "X = pd.concat([X, restricted_df.loc[:, 'France':'Spain']], axis=1)\n",
    "X = pd.DataFrame(preprocessing.MinMaxScaler().fit_transform(X), columns=X.columns) # MinMax Scaling\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 81.4%\n",
      " - Test accuracy: 82.1%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.77      0.92      0.84      2432\n",
      "           1       0.90      0.72      0.80      2432\n",
      "\n",
      "    accuracy                           0.82      4864\n",
      "   macro avg       0.83      0.82      0.82      4864\n",
      "weighted avg       0.83      0.82      0.82      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', max_iter=1000, random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Other models"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Recap: best features (Logistic Regression)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>MScore.2015.int</th>\n",
       "      <th>MScore.2016.int</th>\n",
       "      <th>MScore.2017.int</th>\n",
       "      <th>MScore.2018.int</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120994</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120995</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120998</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121003</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121004</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>24320 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        MScore.2015.int  MScore.2016.int  MScore.2017.int  MScore.2018.int\n",
       "0                     1                0                1                1\n",
       "4                     0                0                1                1\n",
       "7                     0                0                0                0\n",
       "11                    0                0                0                0\n",
       "16                    0                0                0                0\n",
       "...                 ...              ...              ...              ...\n",
       "120994                1                1                0                0\n",
       "120995                1                1                1                1\n",
       "120998                0                0                0                0\n",
       "121003                0                0                0                0\n",
       "121004                1                1                1                1\n",
       "\n",
       "[24320 rows x 4 columns]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for i in range(0, year_widget.value - 2014):\n",
    "    if i == 0: X = restricted_df[['MScore.2015.int']]\n",
    "    else: X = pd.concat([X, restricted_df[['MScore.'+str(2015 + i)+'.int']]], axis=1)\n",
    "y = restricted_df[['MScore.'+str(year_widget.value + 1)+'.int']]\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0, test_size=0.2, shuffle=True, stratify=y)\n",
    "X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Logistic Regression\n",
      " - Train accuracy: 81.5%\n",
      " - Test accuracy: 82.4%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.77      0.91      0.84      2432\n",
      "           1       0.89      0.73      0.81      2432\n",
      "\n",
      "    accuracy                           0.82      4864\n",
      "   macro avg       0.83      0.82      0.82      4864\n",
      "weighted avg       0.83      0.82      0.82      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "lr = LogisticRegression(solver='liblinear', random_state=0)\n",
    "lr.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Logistic Regression', lr, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Features importance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "importance = lr.coef_[0]\n",
    "importance=np.sort(importance)\n",
    "\n",
    "sn.set_style()\n",
    "plt.barh([i for i in range (len(importance))], importance)\n",
    "plt.xlabel('Importance')\n",
    "plt.ylabel('Feature')\n",
    "list_ticks, features_names = [], []\n",
    "for i in range(0, year_widget.value - 2014):\n",
    "    list_ticks.append(i)\n",
    "    features_names.append('MScore '+str(2015 + i))\n",
    "plt.yticks(list_ticks, features_names)\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Decision Tree Classifier\n",
      " - Train accuracy: 81.8%\n",
      " - Test accuracy: 82.2%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.78      0.89      0.83      2432\n",
      "           1       0.87      0.75      0.81      2432\n",
      "\n",
      "    accuracy                           0.82      4864\n",
      "   macro avg       0.83      0.82      0.82      4864\n",
      "weighted avg       0.83      0.82      0.82      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "dtc = DecisionTreeClassifier()\n",
    "dtc.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Decision Tree Classifier', dtc, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(20,20))\n",
    "tree.plot_tree(dtc, label='root', fontsize=9, class_names=['Low', 'High'], filled=True, feature_names=['MScore.2015.int', 'MScore.2016.int','MScore.2017.int','MScore.2018.int', 'MScore.2019.int']);"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest Classifier\n",
      " - Train accuracy: 81.8%\n",
      " - Test accuracy: 82.2%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.78      0.89      0.83      2432\n",
      "           1       0.87      0.75      0.81      2432\n",
      "\n",
      "    accuracy                           0.82      4864\n",
      "   macro avg       0.83      0.82      0.82      4864\n",
      "weighted avg       0.83      0.82      0.82      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "rfc = RandomForestClassifier(n_estimators = 100)\n",
    "rfc.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Random Forest Classifier', rfc, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Support Vector"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Support Vector Classifier\n",
      " - Train accuracy: 81.8%\n",
      " - Test accuracy: 82.2%\n",
      "\n",
      "Test               precision    recall  f1-score   support\n",
      "\n",
      "           0       0.78      0.89      0.83      2432\n",
      "           1       0.87      0.75      0.81      2432\n",
      "\n",
      "    accuracy                           0.82      4864\n",
      "   macro avg       0.83      0.82      0.82      4864\n",
      "weighted avg       0.83      0.82      0.82      4864\n",
      "\n"
     ]
    }
   ],
   "source": [
    "svc = SVC(kernel='poly', random_state=0)\n",
    "svc.fit(X_train, np.ravel(y_train))\n",
    "print_performances('Support Vector Classifier', svc, X_train, y_train, X_test, y_test)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Models comparison through ROC curves"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 720x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(0, figsize=(10, 7)).clf()\n",
    "\n",
    "colors = ['#ff0000', '#ffd700', '#00ff00', '#0000ff', '#800080']\n",
    "\n",
    "# Random Classifier\n",
    "random_fpr, random_tpr, _ = metrics.roc_curve(y_test, [0 for _ in range(len(y_test))])\n",
    "plt.plot(random_fpr, random_tpr, linestyle='--', color=colors[0], label='Random Classifier')\n",
    "\n",
    "# Logistic Regression\n",
    "y_pred = lr.predict(X_test)\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred)\n",
    "plt.plot(fpr, tpr)\n",
    "auc = round(metrics.roc_auc_score(y_test, y_pred), 2)\n",
    "plt.plot(fpr, tpr, color=colors[1], label='Logistic Regression, AUC='+str(auc))\n",
    "\n",
    "# Decision Tree Classifier\n",
    "y_pred = dtc.predict(X_test)\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred)\n",
    "plt.plot(fpr, tpr)\n",
    "auc = round(metrics.roc_auc_score(y_test, y_pred), 2)\n",
    "plt.plot(fpr, tpr, color=colors[2], label='Decision Tree Classifier, AUC='+str(auc))\n",
    "\n",
    "# Random Forest Classifier\n",
    "y_pred = rfc.predict(X_test)\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred)\n",
    "plt.plot(fpr, tpr)\n",
    "auc = round(metrics.roc_auc_score(y_test, y_pred), 2)\n",
    "plt.plot(fpr, tpr, color=colors[3], label='Random Forest Classifier, AUC='+str(auc))\n",
    "\n",
    "# Support Vector Classifier\n",
    "y_pred = svc.predict(X_test)\n",
    "fpr, tpr, _ = metrics.roc_curve(y_test, y_pred)\n",
    "plt.plot(fpr, tpr)\n",
    "auc = round(metrics.roc_auc_score(y_test, y_pred), 2)\n",
    "plt.plot(fpr, tpr, color=colors[4], label='Support Vector Classifier, AUC='+str(auc))\n",
    "\n",
    "plt.ylabel('True Positive Rate (Sensitivity)')\n",
    "plt.xlabel('False Positive Rate (1 - Specificity)')\n",
    "plt.legend()\n",
    "plt.show();"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.2"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}