Experimentation_Phase_1_Data_Exploration.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/aisha-partha/AIMLOps-Assignments/blob/main/Experimentation_Phase_1_Data_Exploration.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Experimentation_Phase : 1. Data_Exploration & Preliminary Model\n"
      ],
      "metadata": {
        "id": "PprHHREe8n5R"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Jw4QHJlfsSF4"
      },
      "source": [
        "## Learning Objectives\n"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "* Understand & explore the data\n",
        "* Perform data preprocessing\n",
        "* Apply  ML algorithms on **Titanic** dataset\n",
        "\n"
      ],
      "metadata": {
        "id": "EsSB95vPsoxS"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Nh70dVHx0G_B"
      },
      "source": [
        "## Dataset Description"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "y-GMJTRb0Iyy"
      },
      "source": [
        "The sinking of the Titanic is one of the most infamous shipwrecks in history.\n",
        "\n",
        "On April 15, 1912, during her maiden voyage, the widely considered “unsinkable” RMS Titanic sank after colliding with an iceberg. Unfortunately, there weren’t enough lifeboats for everyone onboard, resulting in the death of many passengers and crew.\n",
        "\n",
        "While there was some element of luck involved in surviving, it seems some groups of people were more likely to survive than others.\n",
        "\n",
        "[ Data Set Link: Kaggle competition](https://www.kaggle.com/competitions/titanic)\n",
        "\n",
        "<br/>\n",
        "\n",
        "### Data Set Characteristics:\n",
        "\n",
        "**PassengerId:** Id of the passengers\n",
        "\n",
        "**Survived:** Survived or Not, information\n",
        "\n",
        "**Pclass:** Socio-economic status (SES)\n",
        "  * 1st = Upper\n",
        "  * 2nd = Middle\n",
        "  * 3rd = Lower\n",
        "\n",
        "**Name:** Surname, first names of the Passengers\n",
        "\n",
        "**Sex:** Gender of the passengers\n",
        "\n",
        "**Age:** Age of the passengers\n",
        "\n",
        "**SibSp:**\tNo. of siblings/spouse of the passenger aboard the Titanic\n",
        "\n",
        "**Parch:**\tNo. of parents/children of the passenger aboard the Titanic\n",
        "\n",
        "**Ticket:**\tTicket number\n",
        "\n",
        "**Fare:** Passenger fare\n",
        "\n",
        "**Cabin:**\tCabin number\n",
        "\n",
        "**Embarked:** Port of Embarkation\n",
        "  * S = Southampton\n",
        "  * C = Cherbourg\n",
        "  * Q = Queenstown\n"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "![image.png](data:image/png;base64,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)"
      ],
      "metadata": {
        "id": "Vb6tQUILfP4p"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Problem Statement\n",
        "\n",
        "Build a predictive model that answers the question: **“what sort of people were more likely to survive?”** using titanic's passenger data (ie name, age, gender, socio-economic class, etc)."
      ],
      "metadata": {
        "id": "KmusUbENKSEt"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# @title Download the datasets\n",
        "from IPython import get_ipython\n",
        "\n",
        "ipython = get_ipython()\n",
        "\n",
        "notebook=\"U1_MH1_Data_Munging\" #name of the notebook\n",
        "\n",
        "def setup():\n",
        "    from IPython.display import HTML, display\n",
        "    ipython.magic(\"sx wget https://cdn.iiith.talentsprint.com/aiml/Experiment_related_data/titanic.csv\")\n",
        "    ipython.magic(\"sx wget https://cdn.iiith.talentsprint.com/aiml/Experiment_related_data/test_titanic.csv\")\n",
        "    print(\"Data downloaded successfully\")\n",
        "    return\n",
        "\n",
        "setup()"
      ],
      "metadata": {
        "id": "ViFc50xKK-tY",
        "cellView": "form"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!ls"
      ],
      "metadata": {
        "id": "RM8x-pMDLQuq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "List of files present"
      ],
      "metadata": {
        "id": "kLVDFpbSb1ds"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Note:** Use titanic.csv for training & testing purpose and Titanic_Test.csv for submitting the prediction on Kaggle competition."
      ],
      "metadata": {
        "id": "RKzNj9hyNNPJ"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gn6HQH7abkyL"
      },
      "source": [
        "## Import Required Packages"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eiNOMhdK5n-d"
      },
      "source": [
        "import re\n",
        "import numpy as np                # to work with arrays\n",
        "import pandas as pd                 # for data manipulation\n",
        "import matplotlib.pyplot as plt       # for visualization\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "from sklearn.preprocessing import StandardScaler        # for scaling the features\n",
        "from sklearn.ensemble import RandomForestClassifier      # classification model\n",
        "from sklearn.metrics import accuracy_score, f1_score, precision_score, recall_score        # for performance measuring\n",
        "\n",
        "import joblib         # To persist/save the model"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# ========== NEW IMPORTS FOR PIPELINE BUILDING ========\n",
        "\n",
        "# pipeline\n",
        "from sklearn.pipeline import Pipeline\n",
        "\n",
        "# for the preprocessors\n",
        "from sklearn.base import BaseEstimator, TransformerMixin"
      ],
      "metadata": {
        "id": "CDA9ygWSHId4"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "g5fBLdQaJtd4"
      },
      "source": [
        "## 1. Load & Explore the Data Set\n",
        "\n",
        "* Understand different features in the training dataset\n",
        "* Understand the data types of each columns\n",
        "* Notice the columns of missing values\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sW9DMblWkhj_"
      },
      "source": [
        "# Load the dataset\n",
        "data = pd.read_csv(\"titanic.csv\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data.head()"
      ],
      "metadata": {
        "id": "dyDqsnlkdbdM"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "54ktmr8lh8AS"
      },
      "source": [
        "# Getting information about the dataset\n",
        "data.info()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Data Cleaning and Processing\n",
        "## **2. Pre-pipeline Preparation**\n"
      ],
      "metadata": {
        "id": "Gs67b6IN0xHX"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        " ### 2.1 Working on \"SibSp\" & \"Parch\" columns:\n",
        "Combine columns \"SibSp\" & \"Parch\" and create another column that represents the total passengers in one ticket with the name \"family_size\". In each ticket, there might be Siblings/Spouses (SibSp =Number of Siblings/Spouses Aboard) or Parents/Children (Parch=Number of Parents/Children Aboard ) along with the passenger who booked the ticket.\n",
        "\n",
        "  "
      ],
      "metadata": {
        "id": "mUm20kyHVTZe"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Define a function that takes a dataframe as a parameter and returns a transformed\n",
        "# Dataframe\n",
        "def family_size(data_frame):\n",
        "    df = data_frame.copy()\n",
        "    df['FamilySize'] = df['SibSp'] + df['Parch'] + 1\n",
        "    return df"
      ],
      "metadata": {
        "id": "n2vHu13tWJx-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data = family_size(data)\n",
        "data.info()"
      ],
      "metadata": {
        "id": "lEkyLqqEVX2p"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data.head()"
      ],
      "metadata": {
        "id": "R6h2Z-GCzb_4"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-pMXWDyaspe3"
      },
      "source": [
        "### 2.2 Working on \"Cabin\" column:\n",
        "Find unique entries in the Cabin column. We can label all passengers in two categories having a cabin or not. Check the data type(use: type) of each entry of the Cabin. Convert a string data type into '1' i.e. passengers with cabin and others into '0' i.e. passengers without cabin.  Write a function for the above operation and apply it to the cabin column and create another column with the name \" Has_cabin\" containing only 0 or 1 entries.\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Write a function that takes dataframe as a parameter and returns the transformed\n",
        "# Dataframe after binary encoding of the values of 'Cabin' column\n",
        "\n",
        "def process_cabin(data_frame):\n",
        "    df = data_frame.copy()\n",
        "    f1 = lambda x: 0 if type(x) == float else 1         ## Ternary Expression\n",
        "    df['Has_cabin'] = df['Cabin'].apply(f1)\n",
        "    return df"
      ],
      "metadata": {
        "id": "KozjEOo9VS3Y"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data = process_cabin(data)\n",
        "data.info()"
      ],
      "metadata": {
        "id": "n9N82uXWWf81"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data.head()"
      ],
      "metadata": {
        "id": "YEoJzAEGzrTE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 2.3 Working on \"Name\" column :\n",
        "Fetch titles from the name. We can map these titles with numbers and convert them into an integer by mapping with relative numbers. Use: concept of the regular expression."
      ],
      "metadata": {
        "id": "nrMLcBqVh5yE"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Extracts the title (Mr, Ms, etc) from the name variable\n",
        "\n",
        "def get_title(passenger):\n",
        "    line = passenger\n",
        "    if re.search('Mrs', line):\n",
        "        return 'Mrs'\n",
        "    elif re.search('Mr', line):\n",
        "        return 'Mr'\n",
        "    elif re.search('Miss', line):\n",
        "        return 'Miss'\n",
        "    elif re.search('Master', line):\n",
        "        return 'Master'\n",
        "    else:\n",
        "        return 'Other'"
      ],
      "metadata": {
        "id": "w1uZfWxHkDPj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "get_title('heranl Mr.')"
      ],
      "metadata": {
        "id": "VyB7sS5ykHRz"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data['Name'].apply(get_title)"
      ],
      "metadata": {
        "id": "NiiIeAPyjeqY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data['Title'] = data['Name'].apply(get_title)"
      ],
      "metadata": {
        "id": "pPqV3McZh2-j"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data['Title'].unique()"
      ],
      "metadata": {
        "id": "1FBfCUOYoBBW"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data.info()"
      ],
      "metadata": {
        "id": "gH2xsJmW3WXI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data.head()"
      ],
      "metadata": {
        "id": "jVYvAYwn0Pyf"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# drop unnecessary variables\n",
        "\n",
        "data.drop(labels=['PassengerId', 'Name', 'SibSp', 'Parch', 'Ticket', 'Cabin'], axis=1, inplace=True)\n",
        "\n",
        "# display data\n",
        "data.info()"
      ],
      "metadata": {
        "id": "mM2vE18w3x4K"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "data.head()"
      ],
      "metadata": {
        "id": "uIy2mi-00dZI"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 3: Data Exploration\n",
        "### Find numerical and categorical variables"
      ],
      "metadata": {
        "id": "gwMOX7o54i64"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "target = 'Survived'"
      ],
      "metadata": {
        "id": "SjDpu_Ik4zFZ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# List of numerical columns\n",
        "vars_num = [c for c in data.columns if data[c].dtypes!='O' ]#and c!=target]\n",
        "\n",
        "# List of categorical columns\n",
        "vars_cat = [c for c in data.columns if data[c].dtypes=='O']\n",
        "\n",
        "print('Number of numerical variables: {}'.format(len(vars_num)),\":\" , vars_num)\n",
        "\n",
        "print('Number of categorical variables: {}'.format(len(vars_cat)),\":\" , vars_cat)\n"
      ],
      "metadata": {
        "id": "Pjbip-YD46CS"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Find missing values in variables"
      ],
      "metadata": {
        "id": "IYDeF1iI6Ult"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# first in numerical variables\n",
        "data[vars_num].isnull().mean()"
      ],
      "metadata": {
        "id": "MwM5Yxkr6PZO"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# now in categorical variables\n",
        "data[vars_cat].isnull().mean()"
      ],
      "metadata": {
        "id": "8qkwXrOL6aJd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Determine cardinality of categorical variables"
      ],
      "metadata": {
        "id": "mitBWp1h6fv2"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Unique categories in categorical variables\n",
        "data[vars_cat].nunique()"
      ],
      "metadata": {
        "id": "q0-JItKu6gym"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Determine the distribution of numerical variables"
      ],
      "metadata": {
        "id": "aqHFO3Fu6t8v"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Plot a Histogram\n",
        "data[vars_num].hist(bins=30, figsize=(10,10))"
      ],
      "metadata": {
        "id": "G7INSI2j6wBo"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 4: Split the data into train and test set\n",
        "Note: Apply all your data preprocessing steps in train set first and keep test set aside."
      ],
      "metadata": {
        "id": "eQGya6YLOku-"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Features\n",
        "x = data.drop('Survived', axis=1)\n",
        "\n",
        "# Target\n",
        "y = data['Survived']\n",
        "\n",
        "# Split data\n",
        "X_train, X_test, y_train, y_test = train_test_split(x, y, test_size=0.20, random_state=42)\n",
        "\n",
        "X_train.shape, X_test.shape"
      ],
      "metadata": {
        "id": "mps-O7zbPPcL"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## 5. Feature Engineering\n",
        "\n",
        "**Numerical variables:  Age column need to fill.**"
      ],
      "metadata": {
        "id": "2yLdeB-o8kZl"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 5.1 Working on Age column :\n",
        "Find the number of NaN entries in the age column and their row index. Calculate the mean, Standard deviation of the Age column and check the distribution of the age column.We can fill the missing values with randomly generated integer values between (mean+Standard deviation, mean-Standard deviation). Use : np.isnan; np.random.randint; concept of slicing dataframe. Convert the age column as an integer data type.\n",
        "\n"
      ],
      "metadata": {
        "id": "dWmjUnN3VcF6"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "age_avg = X_train['Age'].mean()   # Average Age\n",
        "print(\"Mean of Age :\", age_avg)\n",
        "\n",
        "age_std = X_train['Age'].std()   # Standard Deviation\n",
        "print(\"Std of Age :\",age_std)\n",
        "\n",
        "age_null_count = X_train['Age'].isnull().sum()         # Number of null records in this feature\n",
        "print(\"Count of Nan in Age column :\",age_null_count)\n",
        "\n",
        "## Plotting the histogram plot\n",
        "X_train['Age'].plot(kind='hist',bins=8,edgecolor='r')"
      ],
      "metadata": {
        "id": "gbLqDB1hVf1a"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Define a function, this takes dataframe as a parameter and returns the transformed\n",
        "# Dataframe, filling the missing values of age feature as discussed above.\n",
        "\n",
        "def process_age(data_frame):\n",
        "\n",
        "    df = data_frame.copy()\n",
        "    np.random.seed(0)\n",
        "    age_avg = 29.54\n",
        "    age_std = 14.47\n",
        "    age_null_count = df['Age'].isnull().sum()\n",
        "    age_null_random_list = np.random.randint(age_avg - age_std, age_avg + age_std, size=age_null_count)\n",
        "    df.loc[np.isnan(df['Age']),'Age'] = age_null_random_list\n",
        "    df['Age'] = df['Age'].astype(int)\n",
        "    return df"
      ],
      "metadata": {
        "id": "MFrug-3VakhY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train = process_age(X_train)\n",
        "X_train.info()"
      ],
      "metadata": {
        "id": "a9a69SMpa6aG"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Categorical variables:  Embarked    0.002245 column need to fill.**\n",
        "### 5.2 Working on \"Embarked\" column :\n",
        "The \"embarked\" column represents the port of Embarkation: Cherbourg(C), Queenstown(Q), and  Southampton(S ). Thus, the entries are of three categories in this column. Fill in the missing rows in this column. We can fill it with the most frequent category. Map these categorical string entries into numerical.\n",
        "\n"
      ],
      "metadata": {
        "id": "cGPgnKttVYRL"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(X_train.Embarked.unique())        # Unique Categories\n",
        "print(X_train.Embarked.isnull().sum())          # Number of classes\n",
        "print(X_train.Embarked.mode())          # Most frequent category"
      ],
      "metadata": {
        "id": "IApw-rL1Y1jE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def process_embarked(data_frame):\n",
        "    df = data_frame.copy()\n",
        "    df['Embarked'] = df['Embarked'].fillna('S')\n",
        "    df['Embarked'] = df['Embarked'].map( {'S': 0, 'C': 1, 'Q': 2} ).astype(int)\n",
        "    return df"
      ],
      "metadata": {
        "id": "rVTiJ27OVbzB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train = process_embarked(X_train)\n",
        "X_train.info()"
      ],
      "metadata": {
        "id": "0TFv47f9Y0yO"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train.head()"
      ],
      "metadata": {
        "id": "eZapeIQd1pZJ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **Encodings**"
      ],
      "metadata": {
        "id": "gMvM64L4AiEY"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 5.3 Working on \"sex\" & \"Title\" columns :\n",
        "Map the Sex column as 'female' : 0, 'male': 1, and convert it into an integer data type.\n",
        "\n"
      ],
      "metadata": {
        "id": "doeanDr0VgGV"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_train['Sex'] = X_train['Sex'].map( {'female': 0, 'male': 1} ).astype(int)"
      ],
      "metadata": {
        "id": "dGN92EsEVlTQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Map these categorical string entries into numerical.\n",
        "X_train['Title'] = X_train['Title'].map( {'Mrs': 4, 'Master':3 ,'Miss': 2, 'Mr': 1,'Other':0} ).astype(int)"
      ],
      "metadata": {
        "id": "HeU0RmGUksUA"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X_train.head()"
      ],
      "metadata": {
        "id": "YVYnt5p7ccKK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 5.4 Apply Standard Scalar"
      ],
      "metadata": {
        "id": "S2EfoVojebWn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "scaler = StandardScaler()\n",
        "\n",
        "X_train_scaled = scaler.fit_transform(X_train)\n",
        "\n",
        "X_train_scaled[0,:]"
      ],
      "metadata": {
        "id": "fVWd4PEaeiod"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 5.5 Create a single function for preprocessing the test set (X_test) and apply it.\n",
        "#### **Note**: All the pre-processing steps that were applied on the train set before ML Modelling are also applied on the test set before passing through the predict function."
      ],
      "metadata": {
        "id": "Kwa6Ua9Qgbi7"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "## Create a function\n",
        "\n",
        "def pre_process(data):\n",
        "\n",
        "    df = data.copy()\n",
        "\n",
        "    ## working on Embarked column\n",
        "    df['Embarked'] = df['Embarked'].fillna('S')\n",
        "\n",
        "    ## Working on Age column\n",
        "    np.random.seed(0)\n",
        "    age_avg = 29.54\n",
        "    age_std = 14.47\n",
        "    age_null_count = df['Age'].isnull().sum()\n",
        "    age_null_random_list = np.random.randint(age_avg - age_std, age_avg + age_std, size=age_null_count)\n",
        "    df.loc[np.isnan(df['Age']),'Age'] = age_null_random_list\n",
        "    df['Age'] = df['Age'].astype(int)\n",
        "\n",
        "    ## Encoding\n",
        "    df['Sex'] = df['Sex'].map( {'female': 0, 'male': 1} ).astype(int)\n",
        "    df['Embarked'] = df['Embarked'].map( {'S': 0, 'C': 1, 'Q': 2} ).astype(int)\n",
        "    # Mapping tille categorical string entries into numerical.\n",
        "    df['Title'] = df['Title'].map( {'Mrs': 4, 'Master':3 ,'Miss': 2, 'Mr': 1,'Other':0} ).astype(int)\n",
        "\n",
        "    #df.Fare=df.Fare.fillna(df.Fare.mean())\n",
        "    return df"
      ],
      "metadata": {
        "id": "RLFNNM0SgqZ3"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "## Applyting above function on X_test\n",
        "x_test = pre_process(X_test)\n",
        "\n",
        "x_test.info()\n",
        "x_test.head()\n"
      ],
      "metadata": {
        "id": "urGx3SRcc0kN"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 5.6 Apply standard Scalar transformation to x_test\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "jlA3Gnmrc039"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x_test_scaled = scaler.transform(x_test)\n",
        "x_test_scaled[0,:]"
      ],
      "metadata": {
        "id": "N0SAb9ccc2Qm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Exercise 6. Apply Multiple ML Algo. & display the accuracy\n",
        "#### Expected Accuracy >= 80%\n",
        "\n",
        "#### **Random Forest**\n"
      ],
      "metadata": {
        "id": "zUMFQj-Gc1BO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.ensemble import RandomForestClassifier\n",
        "\n",
        "model_rf = RandomForestClassifier(n_estimators = 150, max_depth = 5,random_state = 42)"
      ],
      "metadata": {
        "id": "1Ajc0XAfdwAf"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Fit the model\n",
        "model_rf.fit(X_train_scaled, y_train)\n",
        "\n",
        "# Predict the model\n",
        "y_pred = model_rf.predict(x_test_scaled)\n",
        "\n",
        "# Calculate the accuracy\n",
        "print(\"Accuracy(in %):\", accuracy_score(y_test, y_pred)*100)\n",
        "print(\"F1-score(in %):\", f1_score(y_test, y_pred)*100)"
      ],
      "metadata": {
        "id": "LL5TXSdxeDTh"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}