TF-IDF BASED AND K-MEAN CLUSTURING BASED  PAGE RANKING ALGORITHMS.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# TF-IDF BASED AND K-MEAN CLUSTURING BASED  PAGE RANKING ALGORITHMS"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### READING DATASET AND SAVING IT INTO A DOCS LIST "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import nltk"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "E:\\dataset\n"
     ]
    }
   ],
   "source": [
    "cd \"\\dataset\" "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_name=os.listdir()\n",
    "\n",
    "docs = [] \n",
    "for i in range(len(dataset_name)):\n",
    "    f=open(dataset_name[i],\"r\");\n",
    "    docs.append(f.read())\n",
    "    f.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Saving queries into querys list"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "querys=['information regarding timeshare resales', 'humans responded respond appearance coyotes urban suburban areas', 'locate past present model railroad layouts', 'methods control type ii diabetes', 'jobs outsourced india', 'describe work home scams promoted internet', 'state kyrgyzstan united states relations', 'relationship seminole indians florida', 'describe origin nature extent spread means controlling kudzu', 'describe evidence heredity play role obesity']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TASK1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Creating term frequency vectors for each document in docs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.feature_extraction.text import CountVectorizer\n",
    "\n",
    "cv = CountVectorizer(stop_words = 'english')\n",
    "\n",
    "vectorized_corpus = cv.fit_transform(docs).toarray()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Saving the vocabulary of the dataset into vocabulary variable."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "vocabulary=cv.vocabulary_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Function to calculate the cosine similarity with all document vectors:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "    This function takes a query vector as in input and calculate the cosine similarity of that query vector to all docs in the dataset and return a list of all cosine similarity of that query with complete documents in docs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "def cos_similarity(vector):\n",
    "    result=[]\n",
    "    for i in range(len(docs)):\n",
    "        dot = np.dot(vectorized_corpus[i],vector)\n",
    "        norma = np.linalg.norm(vectorized_corpus[i])\n",
    "        normb = np.linalg.norm(vector)\n",
    "        cos = dot / (norma * normb)\n",
    "        result.append(cos)\n",
    "    return result   "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Calculating the cosine similarity of each query:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "   * Result_similarities is the list we using to save the result.\n",
    "   * Taking a zero vector \"VECTOR\" of size of vocabulary of dataset\n",
    "   * For every term in the query we increse the frequecy of the term in the vector using the term index in vectorized_corpususing \"vocabulary\"."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "result_similarities=[]\n",
    "for q in querys:\n",
    "    vector=np.zeros(len(vectorized_corpus[0]))\n",
    "    for i in q.split(\" \"):\n",
    "        i=i.lower()\n",
    "        if i in vocabulary:\n",
    "            vector[vocabulary[i]]+=1\n",
    "    result_similarities.append(cos_similarity(vector))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Appending the results with there respective document ID and "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "result_new_similarities=[]\n",
    "for i in range(len(result_similarities)):\n",
    "    doc_dict={}\n",
    "    for j in range(len(result_similarities[i])):\n",
    "        doc_dict[j]=result_similarities[i][j]\n",
    "    result_new_similarities.append(doc_dict)\n",
    "    result_new_similarities[i]=sorted(result_new_similarities[i].items(), key =lambda kv:(kv[1], kv[0]),reverse=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "number_of_relevant_documents=[1,3,1,1,1,1,2,1,1,1] #each element with index i represents number \n",
    "                                                                #of relevant documents present for each i+1th Query\n",
    "relevant_documents_id=[[43], [76, 92, 48], [67], [38], [61], [77], [44, 51], [37], [63], [17]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Query 1 the retrived documents are:\n",
      "\tDocument 93 \tDocument 35 \tDocument 29 \tDocument 43 \tDocument 18 \tDocument 101 \tDocument 6 \tDocument 12 \tDocument 26 \tDocument 100 \tDocument 13 \tDocument 16 \tDocument 2 \tDocument 3 \tDocument 5 \tDocument 14 \tDocument 46 \tDocument 77 \tDocument 39 \tDocument 99 \tDocument 83 \tDocument 51 \tDocument 53 \tDocument 44 \tDocument 58 \tDocument 1 \tDocument 40 \tDocument 50 \tDocument 23 \tDocument 88 \tDocument 97 \tDocument 61 \tDocument 56 \tDocument 92 \tDocument 78 \tDocument 59 \tDocument 91 \tDocument 85 \tDocument 54 \tDocument 19 \tDocument 9 \tDocument 72 \tDocument 60 \tDocument 47 \tDocument 10 \tDocument 15 \tDocument 98 \tDocument 79 \tDocument 48 \tDocument 11 \tDocument 17 \tDocument 96 \tDocument 86 \tDocument 37 \tDocument 76 \tDocument 24 \tDocument 4 \tDocument 68 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.017241379310344827\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 2 the retrived documents are:\n",
      "\tDocument 48 \tDocument 92 \tDocument 76 \tDocument 53 \tDocument 28 \tDocument 46 \tDocument 35 \tDocument 93 \tDocument 73 \tDocument 31 \tDocument 16 \tDocument 63 \tDocument 33 \tDocument 78 \tDocument 54 \tDocument 90 \tDocument 26 \tDocument 37 \tDocument 79 \tDocument 97 \tDocument 58 \tDocument 40 \tDocument 12 \tDocument 11 \tDocument 39 \tDocument 60 \tDocument 8 \tDocument 88 \tDocument 100 \tDocument 99 \tDocument 77 \tDocument 41 \tDocument 42 \tDocument 9 \tDocument 91 \tDocument 96 \tDocument 34 \tDocument 44 \tDocument 2 \tDocument 1 \tDocument 18 \tDocument 43 \tDocument 17 \tDocument 101 \tDocument 4 \n",
      "5th level precision and recall: 0.6 1.0\n",
      "10th level precision and recall: 0.3 1.0\n",
      "15th level precision and recall: 0.2 1.0\n",
      "\n",
      "precision: 0.06666666666666667\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 3 the retrived documents are:\n",
      "\tDocument 67 \tDocument 21 \tDocument 47 \tDocument 7 \tDocument 50 \tDocument 10 \tDocument 100 \tDocument 49 \tDocument 31 \tDocument 74 \tDocument 42 \tDocument 53 \tDocument 44 \tDocument 97 \tDocument 58 \tDocument 84 \tDocument 6 \tDocument 35 \tDocument 81 \tDocument 15 \tDocument 86 \tDocument 24 \tDocument 18 \tDocument 41 \tDocument 9 \tDocument 33 \tDocument 72 \tDocument 99 \tDocument 78 \tDocument 60 \tDocument 37 \tDocument 3 \tDocument 2 \tDocument 82 \tDocument 40 \tDocument 22 \tDocument 8 \tDocument 13 \tDocument 45 \tDocument 79 \tDocument 16 \tDocument 87 \tDocument 48 \tDocument 29 \tDocument 17 \tDocument 14 \tDocument 98 \tDocument 101 \tDocument 4 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.02040816326530612\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 4 the retrived documents are:\n",
      "\tDocument 38 \tDocument 93 \tDocument 85 \tDocument 63 \tDocument 21 \tDocument 96 \tDocument 24 \tDocument 41 \tDocument 79 \tDocument 87 \tDocument 97 \tDocument 3 \tDocument 46 \tDocument 88 \tDocument 98 \tDocument 91 \tDocument 31 \tDocument 70 \tDocument 53 \tDocument 78 \tDocument 45 \tDocument 57 \tDocument 83 \tDocument 48 \tDocument 99 \tDocument 11 \tDocument 33 \tDocument 56 \tDocument 17 \tDocument 16 \tDocument 81 \tDocument 25 \tDocument 27 \tDocument 92 \tDocument 74 \tDocument 72 \tDocument 60 \tDocument 34 \tDocument 73 \tDocument 86 \tDocument 49 \tDocument 76 \tDocument 82 \tDocument 35 \tDocument 101 \tDocument 18 \tDocument 22 \tDocument 15 \tDocument 8 \tDocument 19 \tDocument 29 \tDocument 95 \tDocument 2 \tDocument 14 \tDocument 4 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.01818181818181818\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 5 the retrived documents are:\n",
      "\tDocument 61 \tDocument 8 \tDocument 36 \tDocument 98 \tDocument 44 \tDocument 41 \tDocument 28 \tDocument 26 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.0 0.0\n",
      "15th level precision and recall: 0.0 0.0\n",
      "\n",
      "precision: 0.125\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 6 the retrived documents are:\n",
      "\tDocument 77 \tDocument 2 \tDocument 80 \tDocument 6 \tDocument 26 \tDocument 85 \tDocument 60 \tDocument 49 \tDocument 13 \tDocument 56 \tDocument 27 \tDocument 16 \tDocument 54 \tDocument 50 \tDocument 22 \tDocument 34 \tDocument 14 \tDocument 7 \tDocument 61 \tDocument 63 \tDocument 88 \tDocument 86 \tDocument 41 \tDocument 12 \tDocument 39 \tDocument 83 \tDocument 51 \tDocument 73 \tDocument 70 \tDocument 8 \tDocument 98 \tDocument 1 \tDocument 82 \tDocument 52 \tDocument 78 \tDocument 93 \tDocument 19 \tDocument 42 \tDocument 17 \tDocument 47 \tDocument 59 \tDocument 5 \tDocument 91 \tDocument 44 \tDocument 35 \tDocument 99 \tDocument 9 \tDocument 18 \tDocument 81 \tDocument 15 \tDocument 21 \tDocument 37 \tDocument 92 \tDocument 62 \tDocument 53 \tDocument 3 \tDocument 74 \tDocument 28 \tDocument 101 \tDocument 33 \tDocument 72 \tDocument 58 \tDocument 40 \tDocument 79 \tDocument 97 \tDocument 29 \tDocument 76 \tDocument 4 \tDocument 68 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.014492753623188406\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 7 the retrived documents are:\n",
      "\tDocument 60 \tDocument 51 \tDocument 44 \tDocument 29 \tDocument 16 \tDocument 61 \tDocument 30 \tDocument 42 \tDocument 98 \tDocument 52 \tDocument 101 \tDocument 92 \tDocument 58 \tDocument 39 \tDocument 90 \tDocument 36 \tDocument 27 \tDocument 59 \tDocument 11 \tDocument 93 \tDocument 86 \tDocument 26 \tDocument 43 \tDocument 88 \tDocument 12 \tDocument 77 \tDocument 78 \tDocument 47 \tDocument 97 \tDocument 75 \tDocument 6 \tDocument 23 \tDocument 89 \tDocument 63 \tDocument 25 \tDocument 2 \tDocument 84 \tDocument 54 \tDocument 80 \tDocument 8 \tDocument 22 \tDocument 81 \tDocument 4 \tDocument 19 \tDocument 15 \tDocument 76 \tDocument 10 \tDocument 28 \tDocument 91 \tDocument 100 \tDocument 53 \tDocument 34 \tDocument 73 \tDocument 49 \tDocument 31 \tDocument 87 \tDocument 96 \tDocument 3 \tDocument 74 \tDocument 79 \tDocument 7 \tDocument 33 \tDocument 17 \tDocument 99 \tDocument 46 \tDocument 18 \tDocument 21 \tDocument 48 \tDocument 95 \tDocument 68 \n",
      "5th level precision and recall: 0.4 1.0\n",
      "10th level precision and recall: 0.2 1.0\n",
      "15th level precision and recall: 0.13333333333333333 1.0\n",
      "\n",
      "precision: 0.02857142857142857\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 8 the retrived documents are:\n",
      "\tDocument 37 \tDocument 44 \tDocument 60 \tDocument 101 \tDocument 23 \tDocument 28 \tDocument 13 \tDocument 58 \tDocument 84 \tDocument 17 \tDocument 51 \tDocument 38 \tDocument 33 \tDocument 72 \tDocument 73 \tDocument 3 \tDocument 19 \tDocument 29 \tDocument 81 \tDocument 18 \tDocument 8 \tDocument 99 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.045454545454545456\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 9 the retrived documents are:\n",
      "\tDocument 63 \tDocument 20 \tDocument 80 \tDocument 73 \tDocument 1 \tDocument 40 \tDocument 12 \tDocument 41 \tDocument 8 \tDocument 9 \tDocument 56 \tDocument 53 \tDocument 59 \tDocument 45 \tDocument 6 \tDocument 29 \tDocument 18 \tDocument 101 \tDocument 81 \tDocument 39 \tDocument 17 \tDocument 27 \tDocument 46 \tDocument 87 \tDocument 7 \tDocument 93 \tDocument 33 \tDocument 72 \tDocument 47 \tDocument 44 \tDocument 90 \tDocument 14 \tDocument 25 \tDocument 15 \tDocument 91 \tDocument 92 \tDocument 100 \tDocument 79 \tDocument 16 \tDocument 48 \tDocument 97 \tDocument 11 \tDocument 43 \tDocument 86 \tDocument 3 \tDocument 2 \tDocument 21 \tDocument 99 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.020833333333333332\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 10 the retrived documents are:\n",
      "\tDocument 3 \tDocument 54 \tDocument 87 \tDocument 27 \tDocument 8 \tDocument 41 \tDocument 28 \tDocument 17 \tDocument 74 \tDocument 42 \tDocument 98 \tDocument 9 \tDocument 49 \tDocument 1 \tDocument 84 \tDocument 22 \tDocument 15 \tDocument 96 \tDocument 51 \tDocument 53 \tDocument 7 \tDocument 91 \tDocument 38 \tDocument 33 \tDocument 72 \tDocument 101 \tDocument 73 \tDocument 44 \tDocument 58 \tDocument 82 \tDocument 35 \tDocument 18 \tDocument 92 \tDocument 83 \tDocument 78 \tDocument 88 \tDocument 2 \tDocument 24 \n",
      "5th level precision and recall: 0.0 0.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.02631578947368421\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "precision_with_ranking=[]\n",
    "fifth_level_precision=[]\n",
    "tenth_level_precision=[]\n",
    "fifth_level_recall=[]\n",
    "tenth_level_recall=[]\n",
    "recall_with_ranking=[]\n",
    "k=1\n",
    "for i in result_new_similarities:\n",
    "    ten_precison=0.0\n",
    "    five_precison=0.0\n",
    "    fifteen_precison=0.0\n",
    "    five_recall=0.0\n",
    "    ten_recall=0.0\n",
    "    fifteen_recall=0.0\n",
    "    retrived_documents=0\n",
    "    relevant_retrieved=0\n",
    "    print(\"Query\",k,\"the retrived documents are:\",)\n",
    "    for j in i:\n",
    "        if j[1] != 0 :\n",
    "            print(\"\\tDocument\",j[0]+1,end=\" \")\n",
    "            retrived_documents+=1\n",
    "            \n",
    "            if(j[0]+1 in relevant_documents_id[k-1]):\n",
    "                relevant_retrieved+=1\n",
    "                \n",
    "            if(retrived_documents==5 and number_of_relevant_documents[k-1] != 0 ):\n",
    "                five_precison=relevant_retrieved/5\n",
    "                fifth_level_precision.append(five_precison)\n",
    "                five_recall=relevant_retrieved/number_of_relevant_documents[k-1]\n",
    "                fifth_level_recall.append(five_recall)\n",
    "                \n",
    "            if(retrived_documents==10 and number_of_relevant_documents[k-1] != 0 ):\n",
    "                ten_precison=relevant_retrieved/10\n",
    "                tenth_level_precision.append(ten_precison)\n",
    "                ten_recall=relevant_retrieved/number_of_relevant_documents[k-1]\n",
    "                tenth_level_recall.append(ten_recall)\n",
    "                \n",
    "            if(retrived_documents==15 and number_of_relevant_documents[k-1] != 0 ):\n",
    "                fifteen_precison=relevant_retrieved/15\n",
    "                fifteen_recall=relevant_retrieved/number_of_relevant_documents[k-1]\n",
    "    k+=1\n",
    "    \n",
    "    precision=relevant_retrieved/retrived_documents\n",
    "    recall=relevant_retrieved/number_of_relevant_documents[k-2]\n",
    "    \n",
    "    precision_with_ranking.append(precision)\n",
    "    recall_with_ranking.append(recall)\n",
    "    \n",
    "    print(\"\\n5th level precision and recall:\",end=\" \")\n",
    "    print(five_precison,five_recall)\n",
    "    \n",
    "    print(\"10th level precision and recall:\",end=\" \")\n",
    "    print(ten_precison,ten_recall)\n",
    "    \n",
    "    print(\"15th level precision and recall:\",end=\" \")\n",
    "    print(fifteen_precison,fifteen_recall)\n",
    "    \n",
    "    print(\"\\nprecision:\",precision)\n",
    "    print(\"\\nrecall:\",recall)\n",
    "    \n",
    "    print(\"\\n\\n\\n\\n\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "ASSIGNMENT1 PRESION AND RECALL VALUES"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#assignment1 values of precision and recall\n",
    "fifth_recall_without_ranking=[0.0, 0.3333333333, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0]\n",
    "fifth_precision_without_ranking=[0.0, 0.4, 0.0, 0.0, 0.0, 0.0, 0.0, 0.2, 0.0, 0.0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### COMPARISION OF 5TH LEVEL PRECISION AND RECALL WITH ASSIGNMENT1 VALUES:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "data = [fifth_precision_without_ranking,fifth_level_precision]\n",
    "X = np.arange(10)\n",
    "fig = plt.figure()\n",
    "ax = fig.add_axes([0,0,1,1])\n",
    "ax.bar(X + 0.00, data[0], color = 'b',alpha = 0.8, width = 0.25,label=\"Assignment1\")\n",
    "ax.bar(X + 0.25, data[1], color = 'g',alpha = 0.8, width = 0.25,label=\"Term frequency ranking\")\n",
    "plt.title(\"Comparision of precision bar graph\")\n",
    "plt.xlabel(\"Queries\")\n",
    "plt.ylabel(\"precision scores\")\n",
    "labels=['Q1','Q2','Q3','Q4','Q5','Q6','Q7','Q8','Q9','Q10']\n",
    "ax.set_xticks(X)\n",
    "ax.set_xticklabels(labels)\n",
    "plt.legend()\n",
    "data = [fifth_recall_without_ranking,fifth_level_recall]\n",
    "X = np.arange(10)\n",
    "fig = plt.figure()\n",
    "ax = fig.add_axes([0,0,1,1])\n",
    "ax.bar(X + 0.00, data[0], color = 'b', width = 0.25,label=\"Assignment1\")\n",
    "ax.bar(X + 0.25, data[1], color = 'g', width = 0.25,label=\"Term frequency ranking\")\n",
    "plt.title(\"Comparision of recall bar graph\")\n",
    "plt.xlabel(\"Queries\")\n",
    "plt.ylabel(\"recall scores\")\n",
    "ax.set_xticks(X)\n",
    "ax.set_xticklabels(labels)\n",
    "plt.legend()\n",
    "print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Task2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### creating a tfidf_vectorizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "import nltk\n",
    "\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from nltk.corpus import stopwords\n",
    "sw = set(stopwords.words('english'))\n",
    "from nltk.tokenize import RegexpTokenizer\n",
    "tokenizer = RegexpTokenizer(\"[a-zA-Z@]+\")\n",
    "def filter_words(word_list):\n",
    "    useful_words = [w for w in word_list if w not in sw]\n",
    "    return useful_words\n",
    "def myTokenizer(sentence):\n",
    "    words = tokenizer.tokenize(sentence.lower())\n",
    "    return filter_words(words)\n",
    "tfidf_vectorizer = TfidfVectorizer(tokenizer=myTokenizer)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* passing documents to the tfidf_vectorizer\n",
    "* printing tfidf vector of the complete data\n",
    "* using vocabulary to get the bag of words and there index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "vectorized_corpus_tfidf = tfidf_vectorizer.fit_transform(docs).toarray()\n",
    "vocabulary_tfidf=tfidf_vectorizer.vocabulary_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### But we cannot use this vectorizer because as per the documentation it is not using 1+log(tfidf) for Wtf.\n",
    "#### It uses just raw term frequency and IDF and then normalizing the vectors to calculate TFIDF vectors for every document"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Implementation of TFIDF manualy"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* creating a document frequency[DF] list for each term\n",
    "* Using the results from TFIDF veectorizer from sklearn to get the vocabulary in docs\n",
    "* for this we are using vocabulary_tfidf for finding the index of term"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "idf=[]\n",
    "for i in range(len(vocabulary_tfidf)):\n",
    "    idf.append(0)\n",
    "for w,index in vocabulary_tfidf.items():\n",
    "    for i in vectorized_corpus_tfidf:\n",
    "        if(i[index]>0):\n",
    "            idf[index]+=1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* creating a weighted tfidf function to create \"weighted\" tfidf vectors for the documets passed\n",
    "* using this making tfidf vector for documents and queries also\n",
    "*       Wtd = 1 + log10 (TFt,d), if TFt,d ≥ 1 \n",
    "    * 0, otherwise \n",
    "*       IDFt = log10( N/dft ), if dft ≥ 1 \n",
    "    * 0, otherwise"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "import math\n",
    "import numpy as np\n",
    "def tfidf_vectorizer_function(files):\n",
    "    vectorized_tfidf=[]\n",
    "    for q in files:\n",
    "        vector_tfidf=np.zeros(len(vectorized_corpus_tfidf[0]))\n",
    "        vector_tf=np.zeros(len(vectorized_corpus_tfidf[0]))\n",
    "        for s in q.split(\" \"):\n",
    "            s=s.lower()\n",
    "            if s in vocabulary_tfidf:\n",
    "                vector_tf[vocabulary_tfidf[s]]+=1\n",
    "        for i in range(len(vectorized_corpus_tfidf[0])):\n",
    "            if vector_tf[i] != 0:\n",
    "                vector_tf[i]=1+math.log10(float(vector_tf[i]))\n",
    "        for i in range(len(vectorized_corpus_tfidf[0])):\n",
    "            if idf[i]!= 0:\n",
    "                vector_tfidf[i]=vector_tf[i]*math.log10(len(vectorized_corpus_tfidf)/float(idf[i]))\n",
    "            else:\n",
    "                vector_tfidf[i]=0\n",
    "        vectorized_tfidf.append(vector_tfidf)\n",
    "    return vectorized_tfidf"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* passing the docs and queries to the tfidf vectorizer function and saving them into vectorized_data_docs and vectorized_data_queries respectively"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "vectorized_data_docs=tfidf_vectorizer_function(docs)\n",
    "vectorized_data_queries=tfidf_vectorizer_function(querys)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* defining a function to calculate cosine similarity with the TFIDF vector documents to the passed vector\n",
    "* returns a float value of cosine similarity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cos_similarity_tfidf(vector):\n",
    "    result=[]\n",
    "    for i in range(len(docs)):\n",
    "        dot = np.dot(vectorized_data_docs[i],vector)\n",
    "        norma = np.linalg.norm(vectorized_data_docs[i])\n",
    "        normb = np.linalg.norm(vector)\n",
    "        if normb !=0 and norma !=0:\n",
    "            cos = dot / (norma * normb)\n",
    "        else:\n",
    "            cos=0.0\n",
    "        result.append(cos)\n",
    "    return result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## calculating the cosine similary between queries and documents"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Calculating cosine similarity between each query and all the documents and storing the value in result_similarity_tfidf\n",
    "* Sorting the result_similarity_tfidf with there respective cosine similairty and soting them with there docIDS\n",
    "* This sorted cosine similarity along with there docIDS is stored in result_new_similarities"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "result_similarities_tfidf=[]\n",
    "for vector_tfidf in vectorized_data_queries:\n",
    "    result_similarities_tfidf.append(cos_similarity_tfidf(vector_tfidf))\n",
    "result_new_similarities_tfidf=[]\n",
    "for i in range(len(result_similarities_tfidf)):\n",
    "    doc_dict_tfidf={}\n",
    "    for j in range(len(result_similarities_tfidf[i])):\n",
    "        doc_dict_tfidf[j]=result_similarities_tfidf[i][j]\n",
    "    result_new_similarities_tfidf.append(doc_dict_tfidf)\n",
    "    result_new_similarities_tfidf[i]=sorted(result_new_similarities_tfidf[i].items(), key =lambda kv:(kv[1], kv[0]),reverse=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* retirieving all the documents for each query with similarity greater than zero\n",
    "* as all the documentIDS are sorted accroding to there similarity with querys they are printed in the same order\n",
    "* Calculating the kth level precision and recall for k=5,10,15"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Query 1 the relevant documents are:\n",
      "\tDocument 43 \tDocument 12 \tDocument 100 \tDocument 18 \tDocument 3 \tDocument 24 \tDocument 99 \tDocument 44 \tDocument 29 \tDocument 46 \tDocument 101 \tDocument 86 \tDocument 26 \tDocument 6 \tDocument 39 \tDocument 13 \tDocument 51 \tDocument 85 \tDocument 59 \tDocument 56 \tDocument 50 \tDocument 16 \tDocument 5 \tDocument 35 \tDocument 93 \tDocument 77 \tDocument 54 \tDocument 2 \tDocument 58 \tDocument 40 \tDocument 1 \tDocument 78 \tDocument 61 \tDocument 19 \tDocument 91 \tDocument 92 \tDocument 83 \tDocument 53 \tDocument 14 \tDocument 47 \tDocument 60 \tDocument 88 \tDocument 76 \tDocument 97 \tDocument 4 \tDocument 9 \tDocument 72 \tDocument 11 \tDocument 10 \tDocument 37 \tDocument 96 \tDocument 48 \tDocument 79 \tDocument 15 \tDocument 17 \tDocument 98 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.017857142857142856\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 2 the relevant documents are:\n",
      "\tDocument 48 \tDocument 76 \tDocument 92 \tDocument 53 \tDocument 26 \tDocument 100 \tDocument 93 \tDocument 77 \tDocument 101 \tDocument 11 \tDocument 35 \tDocument 91 \tDocument 34 \tDocument 33 \tDocument 78 \tDocument 9 \tDocument 17 \tDocument 42 \tDocument 46 \tDocument 2 \tDocument 97 \tDocument 4 \tDocument 28 \tDocument 63 \tDocument 90 \tDocument 16 \tDocument 12 \tDocument 58 \tDocument 39 \tDocument 40 \tDocument 37 \tDocument 31 \tDocument 79 \tDocument 60 \tDocument 99 \tDocument 44 \tDocument 41 \tDocument 8 \tDocument 43 \tDocument 18 \tDocument 88 \tDocument 96 \n",
      "5th level precision and recall: 0.6 1.0\n",
      "10th level precision and recall: 0.3 1.0\n",
      "15th level precision and recall: 0.2 1.0\n",
      "\n",
      "precision: 0.07142857142857142\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 3 the relevant documents are:\n",
      "\tDocument 67 \tDocument 50 \tDocument 100 \tDocument 6 \tDocument 2 \tDocument 21 \tDocument 98 \tDocument 18 \tDocument 37 \tDocument 31 \tDocument 24 \tDocument 35 \tDocument 60 \tDocument 40 \tDocument 13 \tDocument 3 \tDocument 72 \tDocument 7 \tDocument 101 \tDocument 47 \tDocument 99 \tDocument 4 \tDocument 42 \tDocument 81 \tDocument 97 \tDocument 33 \tDocument 82 \tDocument 84 \tDocument 53 \tDocument 74 \tDocument 58 \tDocument 22 \tDocument 49 \tDocument 16 \tDocument 44 \tDocument 78 \tDocument 41 \tDocument 86 \tDocument 45 \tDocument 29 \tDocument 87 \tDocument 10 \tDocument 9 \tDocument 48 \tDocument 79 \tDocument 14 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.021739130434782608\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 4 the relevant documents are:\n",
      "\tDocument 38 \tDocument 91 \tDocument 88 \tDocument 98 \tDocument 96 \tDocument 46 \tDocument 70 \tDocument 79 \tDocument 11 \tDocument 85 \tDocument 3 \tDocument 92 \tDocument 97 \tDocument 87 \tDocument 33 \tDocument 4 \tDocument 21 \tDocument 101 \tDocument 78 \tDocument 48 \tDocument 31 \tDocument 24 \tDocument 83 \tDocument 99 \tDocument 53 \tDocument 63 \tDocument 86 \tDocument 34 \tDocument 17 \tDocument 93 \tDocument 56 \tDocument 41 \tDocument 35 \tDocument 81 \tDocument 76 \tDocument 25 \tDocument 18 \tDocument 16 \tDocument 72 \tDocument 45 \tDocument 14 \tDocument 2 \tDocument 27 \tDocument 22 \tDocument 8 \tDocument 82 \tDocument 73 \tDocument 74 \tDocument 19 \tDocument 15 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.02\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 5 the relevant documents are:\n",
      "\tDocument 61 \tDocument 36 \tDocument 26 \tDocument 44 \tDocument 8 \tDocument 98 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.0 0.0\n",
      "15th level precision and recall: 0.0 0.0\n",
      "\n",
      "precision: 0.16666666666666666\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 6 the relevant documents are:\n",
      "\tDocument 77 \tDocument 23 \tDocument 98 \tDocument 37 \tDocument 86 \tDocument 88 \tDocument 99 \tDocument 60 \tDocument 93 \tDocument 70 \tDocument 31 \tDocument 52 \tDocument 39 \tDocument 51 \tDocument 83 \tDocument 5 \tDocument 44 \tDocument 91 \tDocument 26 \tDocument 56 \tDocument 18 \tDocument 50 \tDocument 6 \tDocument 2 \tDocument 101 \tDocument 54 \tDocument 80 \tDocument 27 \tDocument 22 \tDocument 4 \tDocument 16 \tDocument 34 \tDocument 85 \tDocument 42 \tDocument 82 \tDocument 61 \tDocument 62 \tDocument 1 \tDocument 63 \tDocument 13 \tDocument 49 \tDocument 14 \tDocument 59 \tDocument 35 \tDocument 47 \tDocument 17 \tDocument 73 \tDocument 8 \tDocument 9 \tDocument 92 \tDocument 78 \tDocument 81 \tDocument 41 \tDocument 21 \tDocument 28 \tDocument 58 \tDocument 7 \tDocument 19 \tDocument 76 \tDocument 3 \tDocument 29 \tDocument 15 \tDocument 97 \tDocument 53 \tDocument 68 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.015384615384615385\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 7 the relevant documents are:\n",
      "\tDocument 44 \tDocument 51 \tDocument 52 \tDocument 42 \tDocument 39 \tDocument 61 \tDocument 10 \tDocument 93 \tDocument 30 \tDocument 16 \tDocument 92 \tDocument 75 \tDocument 43 \tDocument 29 \tDocument 27 \tDocument 23 \tDocument 11 \tDocument 12 \tDocument 54 \tDocument 98 \tDocument 26 \tDocument 36 \tDocument 6 \tDocument 58 \tDocument 60 \tDocument 19 \tDocument 91 \tDocument 97 \tDocument 101 \tDocument 96 \tDocument 88 \tDocument 25 \tDocument 76 \tDocument 63 \tDocument 15 \tDocument 3 \tDocument 49 \tDocument 78 \tDocument 47 \tDocument 90 \tDocument 81 \tDocument 2 \tDocument 28 \tDocument 8 \tDocument 17 \tDocument 73 \tDocument 89 \tDocument 34 \tDocument 77 \tDocument 59 \tDocument 22 \tDocument 4 \tDocument 84 \tDocument 31 \tDocument 21 \tDocument 86 \tDocument 53 \tDocument 33 \tDocument 87 \tDocument 100 \tDocument 99 \tDocument 74 \tDocument 46 \tDocument 95 \tDocument 18 \tDocument 7 \tDocument 68 \n",
      "5th level precision and recall: 0.4 1.0\n",
      "10th level precision and recall: 0.2 1.0\n",
      "15th level precision and recall: 0.13333333333333333 1.0\n",
      "\n",
      "precision: 0.029850746268656716\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 8 the relevant documents are:\n",
      "\tDocument 37 \tDocument 28 \tDocument 101 \tDocument 81 \tDocument 23 \tDocument 44 \tDocument 58 \tDocument 51 \tDocument 60 \tDocument 73 \tDocument 33 \tDocument 38 \tDocument 17 \tDocument 72 \tDocument 3 \tDocument 99 \tDocument 8 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.058823529411764705\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 9 the relevant documents are:\n",
      "\tDocument 63 \tDocument 20 \tDocument 12 \tDocument 8 \tDocument 23 \tDocument 29 \tDocument 99 \tDocument 46 \tDocument 37 \tDocument 1 \tDocument 40 \tDocument 59 \tDocument 56 \tDocument 45 \tDocument 93 \tDocument 73 \tDocument 101 \tDocument 80 \tDocument 18 \tDocument 86 \tDocument 17 \tDocument 60 \tDocument 31 \tDocument 100 \tDocument 87 \tDocument 53 \tDocument 11 \tDocument 6 \tDocument 90 \tDocument 15 \tDocument 21 \tDocument 27 \tDocument 47 \tDocument 25 \tDocument 14 \tDocument 44 \tDocument 92 \tDocument 72 \tDocument 16 \tDocument 43 \tDocument 7 \tDocument 81 \tDocument 48 \tDocument 2 \tDocument 91 \tDocument 3 \tDocument 9 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.02127659574468085\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "Query 10 the relevant documents are:\n",
      "\tDocument 3 \tDocument 38 \tDocument 23 \tDocument 17 \tDocument 87 \tDocument 8 \tDocument 41 \tDocument 37 \tDocument 99 \tDocument 42 \tDocument 51 \tDocument 60 \tDocument 93 \tDocument 27 \tDocument 74 \tDocument 31 \tDocument 18 \tDocument 1 \tDocument 33 \tDocument 96 \tDocument 9 \tDocument 82 \tDocument 73 \tDocument 54 \tDocument 98 \tDocument 91 \tDocument 53 \tDocument 72 \tDocument 28 \tDocument 7 \tDocument 22 \tDocument 84 \tDocument 24 \tDocument 49 \tDocument 44 \tDocument 92 \tDocument 78 \tDocument 83 \tDocument 2 \tDocument 35 \tDocument 15 \tDocument 88 \tDocument 101 \n",
      "5th level precision and recall: 0.2 1.0\n",
      "10th level precision and recall: 0.1 1.0\n",
      "15th level precision and recall: 0.06666666666666667 1.0\n",
      "\n",
      "precision: 0.023255813953488372\n",
      "\n",
      "recall: 1.0\n",
      "\n",
      "\n",
      "\n",
      "\n",
      "\n"
     ]
    }
   ],
   "source": [
    "k=1\n",
    "number_of_relevant_documents=[1,3,1,1,1,1,2,1,1,1] #each element with index i represents number \n",
    "                                                                #of relevant documents present for each i+1th Query\n",
    "relevant_documents_id=[[43], [76, 92, 48], [67], [38], [61], [77], [44, 51], [37], [63], [17]]\n",
    "precision_with_ranking_tfidf=[]\n",
    "fifth_level_precision_tfidf=[]\n",
    "tenth_level_precision_tfidf=[]\n",
    "fifth_level_recall_tfidf=[]\n",
    "tenth_level_recall_tfidf=[]\n",
    "recall_with_ranking_tfidf=[]\n",
    "for i in result_new_similarities_tfidf:\n",
    "    ten_precison_tfidf=0.0\n",
    "    five_precison_tfidf=0.0\n",
    "    fifteen_precison_tfidf=0.0\n",
    "    five_recall_tfidf=0.0\n",
    "    ten_recall_tfidf=0.0\n",
    "    fifteen_recall_tfidf=0.0\n",
    "    retrived_documents=0\n",
    "    relevant_retrieved=0\n",
    "    print(\"Query\",k,\"the relevant documents are:\",)\n",
    "    for j in i:\n",
    "        if j[1] != 0 :\n",
    "            print(\"\\tDocument\",j[0]+1,end=\" \")\n",
    "            retrived_documents+=1\n",
    "            if(j[0]+1 in relevant_documents_id[k-1]):\n",
    "                relevant_retrieved+=1\n",
    "            \n",
    "            if(retrived_documents==5 and number_of_relevant_documents[k-1] != 0 ):\n",
    "                \n",
    "                five_precison_tfidf=relevant_retrieved/5\n",
    "                \n",
    "                fifth_level_precision_tfidf.append(five_precison_tfidf)\n",
    "                \n",
    "                five_recall_tfidf=relevant_retrieved/number_of_relevant_documents[k-1]\n",
    "                \n",
    "                fifth_level_recall_tfidf.append(five_recall_tfidf)\n",
    "            \n",
    "            if(retrived_documents==10 and number_of_relevant_documents[k-1] != 0 ):\n",
    "                ten_precison_tfidf=relevant_retrieved/10\n",
    "                tenth_level_precision_tfidf.append(ten_precison_tfidf)\n",
    "                ten_recall_tfidf=relevant_retrieved/number_of_relevant_documents[k-1]\n",
    "                tenth_level_recall_tfidf.append(ten_recall_tfidf)\n",
    "            \n",
    "            if(retrived_documents==15 and number_of_relevant_documents[k-1] != 0 ):\n",
    "                fifteen_precison_tfidf=relevant_retrieved/15\n",
    "                fifteen_recall_tfidf=relevant_retrieved/number_of_relevant_documents[k-1]\n",
    "    k+=1\n",
    "    precision_tfidf=relevant_retrieved/retrived_documents\n",
    "    recall_tfidf=relevant_retrieved/number_of_relevant_documents[k-2]\n",
    "    precision_with_ranking_tfidf.append(precision_tfidf)\n",
    "    recall_with_ranking_tfidf.append(recall_tfidf)\n",
    "    \n",
    "    print(\"\\n5th level precision and recall:\",end=\" \")\n",
    "    print(five_precison_tfidf,five_recall_tfidf)\n",
    "    print(\"10th level precision and recall:\",end=\" \")\n",
    "    print(ten_precison_tfidf,ten_recall_tfidf)    \n",
    "    print(\"15th level precision and recall:\",end=\" \")\n",
    "    print(fifteen_precison_tfidf,fifteen_recall_tfidf)\n",
    "    print(\"\\nprecision:\",precision_tfidf)\n",
    "    print(\"\\nrecall:\",recall_tfidf)\n",
    "    print(\"\\n\\n\\n\\n\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Comparision of 5th level precision and recall with task1 data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data = [fifth_level_precision,fifth_level_precision_tfidf]\n",
    "X = np.arange(10)\n",
    "fig = plt.figure()\n",
    "ax = fig.add_axes([0,0,1,1])\n",
    "ax.bar(X + 0.00, data[0],alpha = 0.8, width = 0.25,label=\"Term frequency ranking\")\n",
    "ax.bar(X + 0.25, data[1],alpha = 0.8, width = 0.25,label=\"TFIDF ranking\")\n",
    "plt.title(\"Comparision of precision bar graph\")\n",
    "plt.xlabel(\"Queries\")\n",
    "plt.ylabel(\"precision scores\")\n",
    "labels=['Q1','Q2','Q3','Q4','Q5','Q6','Q7','Q8','Q9','Q10']\n",
    "ax.set_xticks(X)\n",
    "ax.set_xticklabels(labels)\n",
    "plt.legend()\n",
    "data = [fifth_level_recall,fifth_level_recall_tfidf]\n",
    "X = np.arange(10)\n",
    "fig = plt.figure()\n",
    "ax = fig.add_axes([0,0,1,1])\n",
    "ax.bar(X + 0.00, data[0], width = 0.25,label=\"Term frequency ranking\")\n",
    "ax.bar(X + 0.25, data[1], width = 0.25,label=\"TFIDF ranking\")\n",
    "plt.title(\"Comparision of recall bar graph\")\n",
    "plt.xlabel(\"Queries\")\n",
    "plt.ylabel(\"recall scores\")\n",
    "plt.legend()\n",
    "ax.set_xticks(X)\n",
    "ax.set_xticklabels(labels)\n",
    "print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# TASK3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## K-means with Eucledian Distance\n",
    "* k-means with eucleadian distance\n",
    "* assuming k = 30 clusters\n",
    "* assuming maximum iterations 200"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[8, 29, 7, 6, 8, 8, 19, 12, 20, 8, 8, 8, 8, 4, 9, 8, 1, 2, 8, 8, 10, 8, 8, 8, 8, 8, 8, 8, 15, 8, 22, 0, 8, 27, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 25, 8, 17, 8, 8, 8, 8, 14, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 8, 11, 8, 8, 8, 8, 8, 8, 8, 8, 8, 26, 8, 8, 16, 8, 28, 8, 8, 8, 8, 21, 8, 8, 23, 8, 13, 8, 8, 8, 18, 3, 24, 8, 5]\n"
     ]
    }
   ],
   "source": [
    "from sklearn.cluster import KMeans\n",
    "\n",
    "kmeans_model = KMeans(n_clusters=30, max_iter=200,random_state=3425)\n",
    "\n",
    "kmeans_model.fit(vectorized_data_docs)\n",
    "\n",
    "clusters = kmeans_model.labels_.tolist()\n",
    "\n",
    "print(clusters)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## K-means with Cosine Similarity\n",
    "\n",
    "* as Kmeans uses euclidian distance as the metric to make the clusters we normalized the vectors to make it equivalent to cosine similarity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[1, 29, 15, 4, 27, 26, 1, 4, 1, 16, 4, 4, 29, 1, 16, 23, 15, 4, 4, 14, 10, 4, 21, 24, 4, 4, 4, 4, 11, 28, 23, 9, 2, 10, 4, 11, 17, 15, 8, 18, 1, 19, 4, 19, 0, 23, 4, 6, 4, 21, 8, 8, 6, 4, 20, 21, 22, 4, 3, 4, 4, 12, 18, 22, 22, 4, 22, 12, 22, 5, 22, 24, 6, 3, 6, 4, 29, 4, 18, 25, 2, 3, 29, 11, 29, 7, 4, 4, 29, 4, 4, 23, 4, 13, 7, 2, 4, 10, 2, 4, 4]\n"
     ]
    }
   ],
   "source": [
    "from sklearn import preprocessing\n",
    "\n",
    "X_Norm = preprocessing.normalize(vectorized_data_docs)\n",
    "\n",
    "km2 = KMeans(n_clusters=30, max_iter=200, random_state=3425).fit(X_Norm)\n",
    "\n",
    "clus = km2.labels_.tolist()\n",
    "\n",
    "print(clus)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Agglomerative clustering with SINGLE linkage"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from scipy.cluster.hierarchy import dendrogram, linkage\n",
    "\n",
    "Z = linkage(vectorized_data_docs, method='single', metric='euclidean')\n",
    "\n",
    "fig = plt.figure(figsize=(25, 10))\n",
    "\n",
    "dendrogram(Z) \n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Agglomerative clustering with COMPLETE linkage"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "Z = linkage(vectorized_data_docs, method='complete', metric='euclidean')\n",
    "\n",
    "fig = plt.figure(figsize=(25, 10))\n",
    "\n",
    "dendrogram(Z) \n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Agglomerative clustering with AVERAGE linkage"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "Z = linkage(vectorized_data_docs, method='average', metric='euclidean')\n",
    "\n",
    "fig = plt.figure(figsize=(25, 10))\n",
    "\n",
    "dendrogram(Z) \n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Agglomerative clustering with WARD'S METHOD"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "Z = linkage(vectorized_data_docs, method='ward', metric='euclidean')\n",
    "\n",
    "fig = plt.figure(figsize=(25, 10))\n",
    "\n",
    "dendrogram(Z) \n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# TASK 4"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* It will make a list according to label as the name and the list of documents in it"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "def packing_data_into_clusters(data):\n",
    "    from collections import defaultdict\n",
    "    clustered_data=defaultdict(list)\n",
    "    for i in range(len(data)):\n",
    "        clustered_data[data[i]].append(i+1)\n",
    "    clustered_data=sorted(clustered_data.items())\n",
    "    return clustered_data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* It calculates the cosine similarity when two vectors are passed to it.\n",
    "* It returns the cosine similarity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "def cos(vector1,vector2):\n",
    "    dot = np.dot(vector1,vector2)\n",
    "    norma = np.linalg.norm(vector1)\n",
    "    normb = np.linalg.norm(vector2)\n",
    "    if norma!=0 and normb !=0:\n",
    "        cosi = dot / (norma * normb)\n",
    "    else:\n",
    "        cosi = 0\n",
    "    return cosi"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Calculating the average intracluster values\n",
    "* For every cluster there will be a average intracluster value and the average of all the intraclusters is the complete intracluster for a particular method/model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [],
   "source": [
    "def intracluster(cd):\n",
    "    inter_cluster_similarity=[]\n",
    "    for i in cd:\n",
    "        s=0\n",
    "        n=len(i[1])\n",
    "        if n == 1:\n",
    "            inter_cluster_similarity.append(1)\n",
    "        else:\n",
    "            for j in range(n):\n",
    "                for k in range(j+1,n):\n",
    "                    s=s+cos(vectorized_data_docs[i[1][j]-1],vectorized_data_docs[i[1][k]-1])\n",
    "            inter_cluster_similarity.append(s/((n*(n-1))/2))\n",
    "    return sum(inter_cluster_similarity)/len(cd)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Calculating the intercluster values\n",
    "* Intercluster values are calculated using centroids which is defined in the function"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "def centroid(document_ids):\n",
    "    c=np.zeros(len(vectorized_data_docs[0]))\n",
    "    for i in document_ids:\n",
    "        for j in range(len(vectorized_data_docs[i-1])):\n",
    "            c[j]=c[j]+vectorized_data_docs[i-1][j]\n",
    "    for j in range(len(c)):\n",
    "        c[j]=c[j]/len(document_ids)\n",
    "    return c       \n",
    "def intercluster(cd):\n",
    "    centroid_dict={}\n",
    "    for i in cd:\n",
    "        centroid_dict[i[0]]=centroid(i[1])\n",
    "    s=0\n",
    "    for c1,v1 in centroid_dict.items():\n",
    "        for c2,v2 in centroid_dict.items():\n",
    "            s=s+cos(v1,v2)\n",
    "    s=s/len(cd)**2\n",
    "    return s"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plotting the graph of intra-similarity and inter-similarity values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "xaxis=[]\n",
    "for i in range(2,101):\n",
    "    xaxis.append(i)\n",
    "import matplotlib \n",
    "from pylab import rcParams\n",
    "import matplotlib.pyplot as plt\n",
    "    plt.figure(figsize=(25, 10))   \n",
    "    matplotlib.rcParams.update({'font.size': 7})\n",
    "def plotinter_intra(inter,intra):\n",
    "    plt.plot(xaxis,inter)\n",
    "    plt.xticks(xaxis)\n",
    "    plt.ylabel('inter cluster')\n",
    "\n",
    "    plt.show()\n",
    "    plt.plot(xaxis,intra)\n",
    "    plt.xticks(xaxis)\n",
    "    plt.ylabel('intra cluster')\n",
    "\n",
    "    plt.show()\n",
    "    plt.plot(inter,intra)\n",
    "    plt.xlabel(\"inter cluster\")\n",
    "    plt.ylabel('intra cluster')\n",
    "  \n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### The intra-similarity and inter-similarity measures are used as parameters for deciding best K.\n",
    "\n",
    "### Different values of K are calculated starting from minimum and then increasing the value in each trial.\n",
    "\n",
    "### By observing the graph and noticing the point when it starts behaving linearly, the best K is selected.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Best k value for k means euclidean"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### The best K value is : 33"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Plotting of values of K, cluster intra-similarity and inter-similarity combination values.\n",
    "* The graph justifies the answer of selection of the best K."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.cluster import KMeans\n",
    "result=[]\n",
    "intra=[]\n",
    "inter=[]\n",
    "for i in range(2,101):\n",
    "    kmeans_model = KMeans(n_clusters=i, max_iter=50,random_state=3425)\n",
    "    kmeans_model.fit(vectorized_data_docs)\n",
    "    clusters = kmeans_model.labels_.tolist()\n",
    "    clusters = packing_data_into_clusters(clusters)\n",
    "    a=intracluster(clusters)\n",
    "    b=intercluster(clusters)\n",
    "    intra.append(a)\n",
    "    inter.append(b)\n",
    "    #print(\"cluster\",i,\"\\nintracluster similarity=\",a,\"\\nintercluster similarity=\",b)\n",
    "plotinter_intra(inter,intra)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Best K value for kmeans cosine similarity"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### The best K value is : 43"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Plotting of values of K, cluster intra-similarity and inter-similarity combination values.\n",
    "* The graph justifies the answer of selection of the best K."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn import preprocessing \n",
    "result=[]\n",
    "intra=[]\n",
    "inter=[]\n",
    "X_Norm = preprocessing.normalize(vectorized_data_docs)\n",
    "for i in range(2,101):\n",
    "    km2 = KMeans(n_clusters=i, max_iter=50, random_state=3221).fit(X_Norm)\n",
    "    clus = km2.labels_.tolist()\n",
    "    clus = packing_data_into_clusters(clus)\n",
    "    a=intracluster(clus)\n",
    "    b=intercluster(clus)\n",
    "    intra.append(a)\n",
    "    inter.append(b)\n",
    "    #print(\"cluster\",i,\"\\nintracluster similarity=\",a,\"\\nintercluster similarity=\",b)\n",
    "plotinter_intra(inter,intra)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Best K value for Agglomerative Single Linkage"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### The best K value is : 31"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Plotting of values of K, cluster intra-similarity and inter-similarity combination values.\n",
    "* The graph justifies the answer of selection of the best K."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.cluster import AgglomerativeClustering\n",
    "result=[]\n",
    "intra=[]\n",
    "inter=[]\n",
    "for i in range(2,101):\n",
    "    cluster_single = AgglomerativeClustering(n_clusters=i, affinity='euclidean',linkage='single')\n",
    "    cluster_single.fit_predict(vectorized_data_docs)\n",
    "    clusters_single = cluster_single.labels_.tolist()\n",
    "    clusters_single = packing_data_into_clusters(clusters_single)\n",
    "    a=intracluster(clusters_single)\n",
    "    b=intercluster(clusters_single)\n",
    "    intra.append(a)\n",
    "    inter.append(b)\n",
    "    #print(\"cluster\",i,\"\\nintracluster similarity=\",a,\"\\nintercluster similarity=\",b)\n",
    "plotinter_intra(inter,intra)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Best K value for Agglomerative Complete Linkage"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### The best K value is : 28"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Plotting of values of K, cluster intra-similarity and inter-similarity combination values.\n",
    "* The graph justifies the answer of selection of the best K."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "result=[]\n",
    "intra=[]\n",
    "inter=[]\n",
    "for i in range(2,101):\n",
    "    cluster_complete = AgglomerativeClustering(n_clusters=i, affinity='euclidean',linkage='complete')\n",
    "    cluster_complete.fit_predict(vectorized_data_docs)\n",
    "    clusters_complete = cluster_complete.labels_.tolist()\n",
    "    clusters_complete = packing_data_into_clusters(clusters_complete)\n",
    "    a=intracluster(clusters_complete)\n",
    "    b=intercluster(clusters_complete)\n",
    "    intra.append(a)\n",
    "    inter.append(b)\n",
    "    #print(\"cluster\",i,\"\\nintracluster similarity=\",a,\"\\nintercluster similarity=\",b)\n",
    "plotinter_intra(inter,intra)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Best K value for Agglomerative Average Linkage"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### The best K value is : 30"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Plotting of values of K, cluster intra-similarity and inter-similarity combination values.\n",
    "* The graph justifies the answer of selection of the best K."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXkAAAD0CAYAAAB+WlaPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nO3deZwcdZ3/8de3r5nMTGZCMhNykwOEhHBJQAVxxRMBQRFUQFxRibfuz1vx99P19xPPVXcXdEVRFIO68sODXUAERZFDSRTClYAJSUhCkplJ5uqe6fOzf3y/PdMJIUOSnqv6/Xw85tE11XV8q+pb7/p2dXWVMzNERCSaYmNdABERGTkKeRGRCFPIi4hEmEJeRCTCFPIiIhGWGOsCVGptbbX58+ePdTFERCaUVatWdZhZ297eG1chP3/+fFauXDnWxRARmVCccxuf7T2drhERiTCFvIhIhCnkRUQirGoh75xb6Jy7xjl3wx79lzrnVoS/pdWan4iIDK9qIW9m683sHXt560PA+4D3Ah/Y803n3HLn3Ern3Mr29vZqFUdERBid0zUtZtZlZt3A5D3fNLOrzWyZmS1ra9vrFUAiInKARuMSym7nXAtgQO8ozE9ERIKqhbxzbhrwBeAE59yngCVmdgnwr8C/AQ74SrXmJyJyMIolI18shT/fnSsM/Z8rlMiF9wtFo2hGyYxSySiZH79kNvR++N+AkhnZfImBQpFsvgRAIuaIxRypeIxk3JFKxHHOT6dYMqY1pTj72FlVX86qhbyZdQLv3kv/h4F/rNZ8RGR8MDMK5aAs2GAglv+yhaHwHAzSQolCyXcXSruPVyj67kL5vaJRKI9b8t2Vw5T77zlNP355XB+g5ffL88mXSoy3R2kcN3fK+A55ERkdxZKRLRQZyJfIhpZiucU4kC+SLZTCn++XK5bIhv658F6uWNE9+H+xonv3YfMV/fLFoWAeqaB0DpKxGIm4IxlavolYjGTC7dF/6L36pO+XiDkSoZ9/dSTiMZLhNREvt6b33p1KxAank0r4/xOxGPEYxJwj5hzxmMM5iMdCueK+X/l956AuEaM+GacuERvcboMHxfBJwTDiYXrJxMh8RaqQF6kSM6M/XySTK5LJFsnkC/Tniv4v9C939+d990Do9q8l+nNFsoWh4Qby5TD3QT1QKJIvHlyyxhw+vOIxUgkfQkP/D3U3NCR2+z8V9yGbisdDEA6FYqI8bgjXclD60xJD3T48h7qTCR++5ZAtB3c85qq0VcaPRNyRiEN9Mj668x3VuYmMI2ZGtlCiZyBP30CBvmyBvoECvdkC6Wz4P3Sns0X/mivQly2SyRZI54pkcv69TK5Af7643y3bVCJGQyrOpGSc+vA3KelbgM2TktQnY9Qn4tQl49QnY9Ql/Gt9Mk59IrZb/8qWY10iTl0yNhjgdRVhnozrN5C1RCEvE1q+WKK7P09XJk93f47u/jzd/Xl6+gvhNU/PgP+/Z8B39w4U6OnP05ctPKdWcTzmaEzFaapL0FCXoLEuQVNdnCkNKRrr4jSk4jSm/HsNqfhgaDek/P+TQr+G1FCQN6Ti1CfixCLYYpXxRSEv44aZ0Zst0NmXo7MvS2c6R2dfjp3pLDvTef+aydOVybErk6Mrnac3W9jnNCcl47RMStI8KUFzfZK2pjoWtjYxuT7B5Pokk+sTNIfuproETfUJ/xrCfHJ9grpEDOcUxjIxKeRlxOWLJTr6smzvybK9Z4D23iw7erO092Zp7x0Ir1k60jlyhdJep9FUl2BKQ5JpjSkOaUixsLWRKQ2++5DGJC2T/N+UhhTN9YkQ7EmdmpCap5CXg1IqGe19WTbv6mdrVz9Pd/eztWuAp7v72dY9wNPdA7T3ZZ9xrto5mNaYom1yPW2T61g0vYm2pjqmNaWY1lhH6+Q6pjWmmNaUYmpjirrE6H5ZJRIVCnkZViZXYGNnho2daTZ0Zti0M8NT4W9r1wC54u6t78l1CWZOqWdGyySOmtHMoc11HNpSz4zmeqZPrmd6sw/whFrZIiNOIS+APx++tXuAdTv6+PuOPta197G+Pc2THWm29QzsNuwhDUnmTm3g6FktvHrpDOZMmcSs8Df7kEk01yfHaClEZE8K+RrUO5BnzbZeHt3aw5ptPazd1svj2/voq/gSs2VSkoVtjZx6eCsLWhuY39rI/GmNzJvWoBAXmUAU8hGXyRVYvbmb1Zu7WL25m4e3dLOhMzP4/pSGJEceOpnznj+b5x06mcOnN3H49CamNaZ0RYlIBCjkI2ZLVz/3P7mT+zfs5K+buli7rYdS+NJz9pRJHDO7hfNPnMOSWc0sntnMjOZ6hblIhCnkJ7jtPQPc/fcO7lnXyb3rOtnS1Q/4Lz+PnzeFV55+OMfPm8Jxc6YwralujEsrIqNNIT/B5Isl7n9yJ394vJ0/PN7Omm3+Fv1TGpK8aOE0LjttASctmMpRM5ojef8PEdk/CvkJIJMr8Ps17dz26DZ+v2YHPQMFknHHSfOn8snXHMVpR7SyeEazfiIvIs+gkB+ncoUSv1+7g5se3Modj+2gP19kamOKVx09g1csPpQXH9FKU502n4jsm1JinHlkazc/X7mZXz2whV2ZPFMbU5z3/NmcfewsTl4wVadgRGS/KOTHgWyhyM0PPc2192zkwae6SMVjvHLJoZx/4hxefESr7r8iIgdMIT+GujI5fnjPRq67bwMdfTkWtjby2dcu4fUnzGZKQ2qsiyciEaCQHwPtvVm+d9d6fnzfRtK5Ii87ajqXnjqfUxe16stTEakqhfwo6ssWuPqP6/neXesZyBd57XGzeM9LF3HUjOaxLpqIRJRCfhSUSsZP7t/E1297nM50jrOOnclHX3UkC1obx7poIhJxCvkR9sjWbi7/xcM88FQXL1gwle+fuZjj5k4Z62KJSI1QyI+QbKHI1297nO/etZ6pjSm++abjOff4WbpPjIiMKoX8CPj7jl4++JMHePTpHi48eS6fPGMxLQ26Pa+IjD6FfJX99C+b+NxNj9CQSvC9ty7jFUsOHesiiUgNU8hXSb5Y4vM3Pcp1923ktCNa+ZcLjmN6c/1YF0tEapxCvgq6M3nee/0q7v57J+96yUI+fsZRuv2AiIwLCvmDtLWrn7dc82ee2pnhq+cfywXL5o51kUREBinkD8LmXRku/O59dKXzrHjnCzl5wdSxLpKIyG4U8gdoU6cP+N6BPNe98wUcr2vfRWQcUsgfgKe7+3nz1feSzhW5/rIXsnR2y1gXSURkrxTy+ymdLfCOa1fSM1DgZ+96IUfPUsCLyPilG5Xvh2LJ+NBP/8aabT1cedEJCngRGffUkt8PX7z5MW5/bAefP/doXnrk9LEujojIsKoW8s65RuBbQA6408xWhP5vAs4C8sDXzOyxas1zNN304Fa+96cnedsp83nri+aPdXFERJ6Tap6uOQ+4wcwuA86p6P8G4B3Ap4CPVHF+o2Zb9wCX/+IhTpg3hc+ctXisiyMi8pxV83TNHOCh0F2s6P814N+BbcAhe47knFsOLAeYN29eFYtTHaWS8bEbHiRfNL7xxuNJ6HmrIjKBVDOxNuODfrfpmtlfzOzdwI+Bp/YcycyuNrNlZrasra2tisWpjh/du4G7nujgM2cvZr4e8iEiE0w1W/I3Alc6584CbnLOXWdmlzjnzgReCzQBH6vi/EbcuvY+vnjLGk4/so2LTh5/nzJERIZTtZA3szRwaUWvFaH/zcDN1ZrPaLrivx8jFY/x5Tccq4d9iMiEpBPMz+LedZ3csWYH7zl9kW4ZLCITlkJ+L0ol44qbH2NWSz1vP3XBWBdHROSAKeT34qbVW3loSzcfedWR1CfjY10cEZEDppDfw0C+yFduXcuSmc28/oTZY10cEZGDopDfw0//soktXf18+szFxPR0JxGZ4BTyFYol4wf3bOD586bw4iNax7o4IiIHTSFf4c61O9jYmeFSfdkqIhGhkK9w7T0bmNFczxlLZ4x1UUREqkIhHzyxvZe7nujgkhcdRlL3pxGRiFCaBdfes4FUIsaFun2BiESIQh7ozuS58a9beN3xs5jamBrr4oiIVI1CHvj5qqfozxd52yn6wlVEokUhj3/q03FzWlgyq3msiyIiUlU1H/Jbuvp5cHM3ZyydOdZFERGpupoP+Vsf3gagyyZFJJIU8g8/zVEzJrNAT30SkQiq6ZDf0TvAyo271IoXkciq6ZC/7ZHtmMFrdD5eRCKqpkP+1oe3sbC1kecd2jTWRRERGRE1G/K70jnuXd/Jq5fO0PNbRSSyajbkf/vYdool4zU6Hy8iEVazIX/HY9uZ1VLPMbNbxrooIiIjpiZD3sxYuWEXL1w4TadqRCTSajLkN3Rm6EznWDZ/6lgXRURkRNVkyK/csBOAZfMPGeOSiIiMrJoM+VUbd9Fcn+DwNl06KSLRVpMhf/+GnZx42CHEYjofLyLRVnMhvyudY117WufjRaQm1FzIr9q4C4Blh+l8vIhEX82F/MqNu0jGHcfNnTLWRRERGXE1F/KrNu7k6Fkt1CfjY10UEZERV1Mhny0UeXBzt07ViEjNqKmQf3hLD7lCSdfHi0jNqKmQX7XR/wjqxMN0ZY2I1IaaCvmVG3Yxf1oDbZPrxrooIiKjoqZCfs22Xo7WXSdFpIZULeSdc43OuR86577rnLu4ov+ZzrkbnHP/6Zx7VbXmt7+yhSKbd2VYpFsZiEgNedaQd97L92Na5wE3mNllwDkV/U8BLgc+BDxjes655c65lc65le3t7fsxu/2zsTNDyWBRW+OIzUNEZLx51pA3MwNOcs4919b+HOCp0F2s6P8L4Frgl8CKvcznajNbZmbL2tranuOs9t+6HX0AasmLSE1JDPP+qcCvnXP9+Nx/4z6G3YwP+gfY/eDxKeAfQve1wEUHVtSDs74jDcCCVrXkRaR2DBfyFwKLzex+59ycYYa9EbjSOXcWcJNz7jozuyT0/y7ggFsPusQHaN2OPma21NNYN9wii4hEx3CJ900gC9yPb5G/79kGNLM0cGlFrxWh//XA9QdXzIO3riPNQp2PF5EaM9z59l5ge+geGOGyjBgzY/2OPp2PF5GaM1zItwOnOOe+xu5fpk4o7X1ZerMFFup8vIjUmOFO1/wQf049BnSPfHFGxrod/kvXRdPVkheR2jJcS/4dZrbGzB7FX+c+Ia3v8JdPLtTpGhGpMc/aknfO/RxY4pxbChjw6KiVqsrW7UgzKRlnZnP9WBdFRGRUPWvIm9kFzrl5ZrbJOXc40DGK5aqq9R19LGht1IO7RaTmDHe65gPOudnA/wWuHIXyjIj17WmdjxeRmjRcyDcD5wJfBLaOfHGqbyBf5KldGV1ZIyI1abiQvxNoMLPVwBMjX5zq29iZwUxX1ohIbRou5FNAu3Purfhfvk4469rDlTVqyYtIDRou5LeFvz7gtJEvTvWtL4e8bmkgIjVonz+GMrPflLudc0eOfHGqb117mlkt9TSkdGMyEak9+0y+cK28ASVg9aiUqMrWt/fpR1AiUrOGa8lfMFoFGSmbd/XzqqP1XFcRqU3D/eLVKvsN89CQcadQLLEzk2P65LqxLoqIyJgY7hevcTMrAjjn4qNXrOrYmc5hBq0KeRGpUcNdXfPViu4vjGRBRsKOXn/VZ1uTQl5EatNwId9Q0T3hvr3s6AshPzk1xiURERkbw11XeLdz7kb81TU3jUJ5qqp9sCWvu0+KSG0a7uqa64DrRqksVdfRlwOgVS15EalRw52umdDae7M0puL6IZSI1Kxoh3xfljZdWSMiNWy4X7yeC1yEv+Vw1sxeNyqlqpKOXoW8iNS24VryZ+Af+3cucN/IF6e62vuytOrySRGpYcOFfCfQCJwMTLgblLWrJS8iNW64bySvBvqBS4DvjHxxqidbKNLdn9cPoUSkpg3Xkn+zmbWb2dfNbEKdrukcvHxSIS8itWu4lvw5zrlFQDeAmX185ItUHe26pYGIyLAh/y78U6Fggt3WYOiWBgp5Ealdw52ueYOZbTSzjcDFo1Ggaim35HW6RkRq2XD3k1/inFsaem0fnSJVx2DIN+mWBiJSu4a7n/zRZvbIaBaoWjr6srRMSlKXmHC3wRcRqZrhzsmf65z7NOAAzOyikS9SdfgfQqkVLyK1bbiQd2Y2oc7Fl+mHUCIiw4f8IufcBUAawMxuHvkiVUdHX46jZzWPdTFERMbUcCH/B/zToRrY46Hee3LONQLfAnLAnWa2IvR/B3ASMBdYbWafOthCPxdqyYuI7OMSSufcqUB7+OsIf/tyHnCDmV0GnFPuaWbXmNm7gceBaw+2wM9Ff65IX7agkBeRmrevlvzU8Fe2z5Y8MAd4KHQXK99wztUDC8xs7Z4jOeeWA8sB5s2bN1x5n5PyD6F0B0oRqXX7uoRyf5/puhkf9A/wzE8I5wM3Pst8rsbfCI1ly5YNdyB5Tnb06teuIiJQ3SdD3Qi8wTn3beAm51zls2EvAH5exXnt0+AtDdSSF5EaV7WHn5pZGri0oteKivfOrdZ8not2teRFRICIPuO1vTeLczC1UT+GEpHaFsmQ7+jLMrUhRTIeycUTEXnOIpmC7b16tquICEQ15Pv0QygREYhoyHco5EVEgKiGfG+OafrSVUQkeiFfKJbozxdpnpQc66KIiIy5yIV8Ju/vqNCQ0sNCRESiF/JZH/KNdVX7nZeIyIQVuZBP5wqAWvIiIhDBkB9syafUkhcRiVzIqyUvIjIkciGfKYe8zsmLiEQx5Muna9SSFxGJXsiHc/JqyYuIRDDky+fk1ZIXEYlgyJdP1zTo6hoRkeiFfDpbIBFzpBKRWzQRkf0WuSTM5Iq6fFJEJIhcyKezBd3SQEQkiFzIZ/JqyYuIlEUv5NWSFxEZFLmQT+ucvIjIoMiFfCZX0OWTIiJB9EI+q5a8iEhZ5EI+nSvoNsMiIkHkQj6TK9JQp5a8iAhELOTNjEyuqJa8iEgQqZDPFkoUS6aWvIhIEKmQH7w5WVIhLyICEQv5dFZPhRIRqRSpkB96KpRCXkQEIhfy5Za8TteIiEDkQl4teRGRSpEK+cFz8vrFq4gIELGQH3r0n0JeRASgauc1nHONwLeAHHCnma0I/WcCnwIc8FMzu7ta89zT4EO8dXWNiAhQ3Zb8ecANZnYZcE5F/48CvUAJ2LznSM655c65lc65le3t7QdVgExWLXkRkUrVDPk5wFOhu1jR/2jgR8DngP+950hmdrWZLTOzZW1tbQdVgHJLXrcaFhHxqhnym/FBv+d0NwO7gD6gvorze4b+XJH6ZIx4zI3kbEREJoxqNnlvBK50zp0F3OScu87MLgG+DnwFMODbVZzfM+g2wyIiu6taIppZGri0oteK0P9R4G3Vms++ZLJFJul8vIjIoEhdQqmWvIjI7iIV8npgiIjI7iIV8umsWvIiIpUiFfKZnB7iLSJSKXIhr1+7iogMiVjIF3R1jYhIhUiFfDpbpFEhLyIyKDIhXywZ/fmibmkgIlIhMiHfnw8PDNEllCIigyIT8hndnExE5BmiE/JZteRFRPYUmZAv32Z4UlIteRGRssiE/OBDvNWSFxEZFJmQH3qIt1ryIiJlkQl5teRFRJ4pMiFfbsnrBmUiIkMiE/Ll6+R1gzIRkSGRCfl0thzyasmLiJRFJuQzuQLOQX0yMoskInLQIpOI/uZkCZxzY10UEZFxIzIhn8kVdD5eRGQPkQn5tB4YIiLyDJEJ+X615EVEniEyIZ/O6vmuIiJ7ikzI+3PyOl0jIlIpMiHvz8mrJS8iUikyIZ/JqiUvIrKnyIR8OqeHeIuI7CkyId+fK9KgSyhFRHYTiZDPFUrkiiUakmrJi4hUikTI94d7yaslLyKyu0iEPMBZx87k8OlNY10MEZFxJRJN35aGJFdd9PyxLoaIyLgTmZa8iIg8k0JeRCTCqna6xjnXCHwLyAF3mtmK0P9zwGJgF/B5M9tarXmKiMi+VbMlfx5wg5ldBpxT0b+AD/480FXF+YmIyDCqGfJzgKdCd7Gi/xVmdgnwW+Cde47knFvunFvpnFvZ3t5exeKIiEg1Q34zPuh3m66ZlULnDuAZ1zia2dVmtszMlrW1tVWxOCIiUs1LKG8ErnTOnQXc5Jy7zswucc59GpgLtAIfrOL8RERkGFULeTNLA5dW9FoR+l9RrXmIiMj+cWY21mUY5JxrBzYexCRagY4R7B6NeYyXeWtZozlvLevEmN/+OszM9n6+28wi8wesHMnu0ZjHeJm3ljWa89ayToz5VfNPP4YSEYkwhbyISIRFLeSvHuHu0ZjHeJn3aM9PyxrN+dXSslZ7WlUxrr54FRGR6opaS15ERCoo5EVEIiwSDw1xzr0OOAuYDvwaOAl/zekdwI+APwI3AGcCjwA/A14BNAMrgfXAxfj1cTywFX+9ahdwJLAwdF8MXAMcBjyA/yXv4cCFwL8BKWAVMC/M/+/ANvxdOIvALPzN2n4B/AX4HJAE3oI/H9cEDITpJIDtoV8T0A5YmO5OIA4cAXw3LPsk/L2DGoFDgB7gz8CLwnhtQDqUqRTG3QZMBerCPJ8AZgCTgSuA9wL1oRyZsB56QhmPBH4CvCwM0xX6HwpkgXuBY0J5Y/gb1G3H395iKf73ENvC+p4F/FeYZhvQDdwNvBG4P0zrA2HdPoT/5fQHgH8AXhO275PAW4FH8TfIexjYEMp2GHBzmMdVwD34u6KeHcr1O/yN9M4Fvgm8O6zjHWHcTCjTXHyd+Xdgeei3BV8H0mG4TNh29cAiYFNYlg3A+/H16sfA/wrzmB62T2dYPy1h/C1hfR4e1m0urLd/DeslGcbPVazz34f1OTtMox9/u5ErgG+Efj8A3hS2cXNYhgS+PmXx9aANWA3MDNNeG8qxE193YmGYtcCUsAxvAT4etmFHWI40fvvfja9vh4ZtHg/rtXyvq+lhe+TC+nNAQ/g/Ht7/JfCqsE4GKspcDMtzXZj+tcDbw/S6w/adWlGWbChzPqwfF9Z/ISxfCr+/9YVtOBO/z78pjNuD31+SYZgP4ff92cBXgE+E6WwPy9YWtu3ssDwtoX9TKPuPw3adCfwNX69gL3f0PVCRaMmb2S/N3/3ybcApZvZu/I6wDL/S/xO/Icsb7ij8Ss8Dm83srjDOfwF3Af9tZm8HXgJ8ycwW439ssB5fgf8KfB+/cZ8O03ge/oCxGX9HzlX4YLkeX3Gewge7w1eANwI34SvXA2Z2HH4jP4k/gKzC79A9wAmh7HngRPzG/yD+gDY1zPtYfEU5ATguLO/P8IHUgD+43Y2voIR5p8zsKOAFoRzt4f/vAS8H7sPvnAl8XTkcH3xXhWWfEYZfiK/0F+CDrSus85+HctyPP+Amwnz+PzANeLuZHR3KlccH/yL8DnoO8DV8qF8G/As+KKfjA3wX/uD8FfyO1Ad8Fb9DXRmmCX7H3hCG/xL+INAbtsc38CFcF17vwx9wCqHM2xhqCDzO0MF/Pb4O3IM/uD4cxn06bOs8Pohy4f91wOtCvwK+jpS38Ub8zv5nfOgcgr+ZXyxsl9vCMIkwnXIY3IoPuQ7gN/g6dWzYDoUwvIXl+nJYxmxY74+HeQyE5bs1bKs+4JNh+zSE7hVhe98O/CRs74+H5b8gDHN7mNfN+Pp7ZVh/m4A/hPKsDf1OBD6Fb2ysxu+f9wIWph0HppjZkWE93hemtwT4QqjrLcCPwjBp/MF6XdjmJ+IPkH/Eh/kkfF2fBnytYh47wrRiQNbMluAPPE2h+1DgV2HZzwfeGsZdhL9l+pKwjt4elusefD1dG8qxJEz7ZuBo4F2hrjcDT5jZlPD+SfiDwufw9ec8nv2OvgckEiFf4TPAVc65c4A/4Sv+o+H1ETN7Db5SLQfuNbMPA++pGP8ifEvuzc653+E/BbzZOfdV/M4KvjKD3/HmsLvJ+J1pO76ynQW8Ht9CBr9T/BW/o54EXI4PpYvD+4eE93+LP8C8G98y+V0YzsIy/izM+1iGvpH/Nn5nm46vZO34lunh+B3g5aFcuTCfAnC7c+5l+JbKD4C1zrlbQnm+GKa7GR8Gs8I6vQN4Jb5y3xfW9Sp8BT0pvDYC/4QPth1hPS7FB9JLgNPw4XWxc+4j+IPS7aF8u/AHjVvwn7Zejm/dlT/VlD9JHIPfti34MNhSsbxtYfwWfKBchW8FWlifLfiD7Wx8iGTxO9MM/E75jbCeXoo/MH8CWIA/6K4Ezgh16cawrJeGsh2H3+HPD+urDXgx8GEAM1uIP/i9PYx/Vyj7NfhPCaeE6Zd3/kyooyfht/9vwnLcG7b784A78cFxGr4B84owrS+Z2bHA6fhPkt8CPh+2273AT/EheBe+9f+usG6+im+tT8LvDz/G17VjwrohDLshbJOvh+V8PX7fmYv/hHcBviW9LZS/GJbzYvwnmpvx9fw7+JZ23jn3FL6B0Bb2v1vwn9b+Ef8p5c3OufvwB48znHNbw/o/Ch/iL8E3fLrx9WALPmBPDuvu/c65+8M2anTOlT+ZdDjn/oo/QE4P+3tPWJevxB983+ycuyYsxyucc6vDeqnD16Vj8AecR8L8DZ8BcXyIX+6cuy5MF+fcRnyduTas/9PCsHN49jv6HpBIhLzzvgzcYmZ/NbNfm9kp+Ir2QnxlfadzLoYPkXx4hbASnXPz8BvnAuCzZvYy4CVm9j58ayW3x2zn4QOwXIal+J3ug2aWwwffr/A7/rHh9Yww+A5gk5kVQlkmh/4L8J8Qzsa3RC7Hf0z8GD4U5+ErfiEMswV40Dl3Pb6CLMefBvggsAZ/KmM+voXywzDuevxOdQu+JfU+/M7+fXwwrca36mab2TvMbBb+oLQprNM34HfOa/AHsfVhWXNmdqOZHRqWaQq+dboUyIdxFwJrQ/epYbmPxFfwi4F5ZtYSluVI/I4yD78zXIQPuJlhPV2ED9OL8GG9GN8SnR3G/efwem7FMPHQfRJ+xzofHxBL8WGZAD6ND6v34wPH8HWlJWzvHJByzr0UH6BrzWxbmG4/Plwuxx9ITw7jpsM2J2yzSaH7zDDOrLCtOoAnzeyfwjh9YbgSQ3WtP6yPj+Jbwx1m9k58mM7Af7JJhW1eHn47/iBRWfffGcqyONT11aEcd+JDMIvfH6aFacdO7S0AAAL3SURBVAwAaefcK8L6f0+Y3x34g91x+IZGM/7ANieMvznMvwv/SWhu2Bb/gd/Xfo0/sN1mZnPxdWR9KNNF+APGD8O2uhb/qakF+D+hbvbgD8btYX4fYWj/XRy2z+n4sP4t/lPQgrDepoby/YqhBsoa/P6exh8Ybw/T/nZYn/eZ2Xvwn5jX4+tYFp+lxYq82IX/tPaJsPy/Duuw/EmyfAq4IZTpgbC9N/Msd/Q9UJG4hNI590H80f5+hs6J1gGrzewq59zb8KcdWhkKtDeGYdeEYf4Zf7TvwX906sCv4Dp8C6shjPcifIV4El+pXoBvKS/E70BPhHHm4lv7f8NX+iw+ZGbgP9J9B9/Cm4dvlU3BB/o38AG+EF/RYqEsKXw47AplqcOHZA9+53wc35KbjN/hG/GV6C/4Vmr5Ewb4kOgKr+VzoFcCl+BDqg9fuY8P0+kNw88Nwz6NP3DtDP02MXQQjIXl/BM+nF6GPyiVGAqreWEe3wD+Hz4o+sN6agjr6rf4llAGHzyvDtO5HX/QexzfSnsoDP9kKDv4YH8j/juFNWF+7fhW69lhGT6A33G/HobP4UP57lCOlrBc9fh60xXmvxTfij0CH5JdYbskw3p6gKFWfXld/DksxwvC+vwRvkFwcui+kKHvYKbgDyxbwjZrDesmVzHPxWF5d+EP2Dl83VsbtslpYT4D+PrxH/iDl8O3zJ+Hr9NX4j+FxMKylT8xbcEHMWH8aWHdrg7r7/dhO54Y1kUWH67vxTeMTnfOPYAPwU78AS+Gr8+dwHFhmD+F7dIUtlE+DNcb+qXC/OeGcWeF5Sas70xYv+8J2/MEfOPo9DB8Iqyj8ifsx/EH9T/g9+Ni2HY9+Hr397C954R1EMfXsbVh2lvx+3r5O48N+AP6cnwD7sP47z/A7/suLNfTYXrlT3WfDctUh98/c/hcWIVvdBG2zQDwp4M9Jx+JkBcRkb2LxOkaERHZO4W8iEiEKeRFRCJMIS8iEmEKeRGRCFPIi4hE2P8AFFBz87IvPcgAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.cluster import AgglomerativeClustering\n",
    "result=[]\n",
    "intra=[]\n",
    "inter=[]\n",
    "for i in range(2,101):\n",
    "    cluster_average = AgglomerativeClustering(n_clusters=i, affinity='euclidean',linkage='average')\n",
    "    cluster_average.fit_predict(vectorized_data_docs)\n",
    "    clusters_average = cluster_average.labels_.tolist()\n",
    "    clusters_average = packing_data_into_clusters(clusters_average)\n",
    "    a=intracluster(clusters_average)\n",
    "    b=intercluster(clusters_average)\n",
    "    intra.append(a)\n",
    "    inter.append(b)\n",
    "    #print(\"cluster\",i,\"\\nintracluster similarity=\",a,\"\\nintercluster similarity=\",b)\n",
    "plotinter_intra(inter,intra)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Best K value for Agglomerative Ward's Method"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### The best K value is : 37"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Plotting of values of K, cluster intra-similarity and inter-similarity combination values.\n",
    "* The graph justifies the answer of selection of the best K."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1800x720 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.cluster import AgglomerativeClustering\n",
    "result=[]\n",
    "intra=[]\n",
    "inter=[]\n",
    "for i in range(2,101):\n",
    "    cluster_ward = AgglomerativeClustering(n_clusters=i, affinity='euclidean',linkage='ward')\n",
    "    cluster_ward.fit_predict(vectorized_data_docs)\n",
    "    clusters_ward = cluster_ward.labels_.tolist()\n",
    "    clusters_ward = packing_data_into_clusters(clusters_ward)\n",
    "    a=intracluster(clusters_ward)\n",
    "    b=intercluster(clusters_ward)\n",
    "    intra.append(a)\n",
    "    inter.append(b)\n",
    "    #print(\"cluster\",i,\"\\nintracluster similarity=\",a,\"\\nintercluster similarity=\",b)\n",
    "plotinter_intra(inter,intra)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TASK 5"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Computation of precision and recall for each query is done on the dataset which is in the form of clusters"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "* Plotting function for plotting precision and recall"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_precision_recall(precision,recall,s):\n",
    "    data = [precision_with_ranking_tfidf,precision]\n",
    "    X = np.arange(10)\n",
    "    fig = plt.figure()\n",
    "    ax = fig.add_axes([0,0,1,1])\n",
    "    ax.bar(X + 0.00, data[0], color = 'b',alpha = 0.8, width = 0.25,label=\"TFIDF ranking\")\n",
    "    ax.bar(X + 0.25, data[1], color = 'g',alpha = 0.8, width = 0.25,label=s)\n",
    "    plt.title(\"Comparision of precision bar graph\")\n",
    "    plt.xlabel(\"Queries\")\n",
    "    plt.ylabel(\"precision scores\")\n",
    "    labels=['Q1','Q2','Q3','Q4','Q5','Q6','Q7','Q8','Q9','Q10']\n",
    "    ax.set_xticks(X)\n",
    "    ax.set_xticklabels(labels)\n",
    "    plt.legend()\n",
    "    data = [recall_with_ranking_tfidf,recall]\n",
    "    X = np.arange(10)\n",
    "    fig = plt.figure()\n",
    "    ax = fig.add_axes([0,0,1,1])\n",
    "    ax.bar(X + 0.00, data[0], color = 'b',alpha = 0.8, width = 0.25,label=\"TFIDF ranking\")\n",
    "    ax.bar(X + 0.25, data[1], color = 'g',alpha = 0.8, width = 0.25,label=s)\n",
    "    plt.title(\"Comparision of recall bar graph\")\n",
    "    plt.xlabel(\"Queries\")\n",
    "    plt.ylabel(\"recall scores\")\n",
    "    labels=['Q1','Q2','Q3','Q4','Q5','Q6','Q7','Q8','Q9','Q10']\n",
    "    ax.set_xticks(X)\n",
    "    ax.set_xticklabels(labels)\n",
    "    plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Queries on kmeans with euclidean distance\n",
    "#### Precison and Recall for each query is calculated and the results are compared with the Precision and Recall values of Task 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "kmeans_model = KMeans(n_clusters=33, max_iter=200,random_state=3425)\n",
    "kmeans_model.fit(vectorized_data_docs)\n",
    "cluster1 = kmeans_model.labels_.tolist()\n",
    "clus1=packing_data_into_clusters(cluster1)\n",
    "count=1\n",
    "precision=[]\n",
    "recall=[]\n",
    "for q in querys:\n",
    "    x=tfidf_vectorizer.transform([q])\n",
    "    pred = kmeans_model.predict(x.toarray())\n",
    "    for i in clus1:\n",
    "        if i[0] ==pred:            \n",
    "            retrieved=len(i[1])\n",
    "            relevant_retrived=0\n",
    "            for j in relevant_documents_id[count-1]:\n",
    "                if j in i[1]:\n",
    "                    relevant_retrived+=1\n",
    "    precision.append(relevant_retrived/retrieved)\n",
    "    recall.append(relevant_retrived/len(relevant_documents_id[count-1]))\n",
    "    count+=1\n",
    "plot_precision_recall(precision,recall,\"kmean_euclidian\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Queries on kmeans with cosine similarity\n",
    "#### Precison and Recall for each query is calculated and the results are compared with the Precision and Recall values of Task 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "km2 = KMeans(n_clusters=43, max_iter=200, random_state=3221).fit(X_Norm)\n",
    "clus = km2.labels_.tolist()\n",
    "clus = kmeans_model.labels_.tolist()\n",
    "clus1=packing_data_into_clusters(clus)\n",
    "count=1\n",
    "precision=[]\n",
    "recall=[]\n",
    "for q in querys:\n",
    "    x=tfidf_vectorizer.transform([q])\n",
    "    pred = kmeans_model.predict(x.toarray())\n",
    "    for i in clus1:\n",
    "        if i[0] ==pred:            \n",
    "            retrieved=len(i[1])\n",
    "            relevant_retrived=0\n",
    "            for j in relevant_documents_id[count-1]:\n",
    "                if j in i[1]:\n",
    "                    relevant_retrived+=1\n",
    "    precision.append(relevant_retrived/retrieved)\n",
    "    recall.append(relevant_retrived/len(relevant_documents_id[count-1]))\n",
    "    count+=1\n",
    "plot_precision_recall(precision,recall,\"kmean_cos_similarity\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Queries on Agglomerative Clustering with Single linkage\n",
    "#### Precison and Recall for each query is calculated and the results are compared with the Precision and Recall values of Task 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.cluster import AgglomerativeClustering\n",
    "cluster_single = AgglomerativeClustering(n_clusters=31, affinity='euclidean',linkage='single')\n",
    "cluster_single.fit_predict(vectorized_data_docs)\n",
    "clusters_single = cluster_single.labels_.tolist()\n",
    "clusters_single_dict=packing_data_into_clusters(clusters_single)\n",
    "centroid_single=[]\n",
    "precision=[]\n",
    "recall=[]\n",
    "for i in clusters_single_dict:\n",
    "    centroid_single.append(centroid(i[1]))\n",
    "count=1\n",
    "for q in vectorized_data_queries:\n",
    "    cos_single=[]\n",
    "    for i in range(len(centroid_single)):\n",
    "        cos_single.append(cos(centroid_single[i],q))\n",
    "    index=cos_single.index(max(cos_single))\n",
    "    for i in clusters_single_dict:\n",
    "        if i[0] == index:            \n",
    "            retrieved=len(i[1])\n",
    "            relevant_retrived=0\n",
    "            for j in relevant_documents_id[count-1]:\n",
    "                if j in i[1]:\n",
    "                    relevant_retrived+=1\n",
    "    precision.append(relevant_retrived/retrieved)\n",
    "    recall.append(relevant_retrived/len(relevant_documents_id[count-1]))\n",
    "    count+=1\n",
    "plot_precision_recall(precision,recall,\"single\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Queries on Agglomerative Clustering with Complete linkage\n",
    "#### Precison and Recall for each query is calculated and the results are compared with the Precision and Recall values of Task 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cluster_single = AgglomerativeClustering(n_clusters=28, affinity='euclidean',linkage='complete')\n",
    "cluster_single.fit_predict(vectorized_data_docs)\n",
    "clusters_single = cluster_single.labels_.tolist()\n",
    "clusters_single_dict=packing_data_into_clusters(clusters_single)\n",
    "centroid_single=[]\n",
    "precision=[]\n",
    "recall=[]\n",
    "for i in clusters_single_dict:\n",
    "    centroid_single.append(centroid(i[1]))\n",
    "count=1\n",
    "for q in vectorized_data_queries:\n",
    "    cos_single=[]\n",
    "    for i in range(len(centroid_single)):\n",
    "        cos_single.append(cos(centroid_single[i],q))\n",
    "    index=cos_single.index(max(cos_single))\n",
    "    for i in clusters_single_dict:\n",
    "        if i[0] == index:            \n",
    "            retrieved=len(i[1])\n",
    "            relevant_retrived=0\n",
    "            for j in relevant_documents_id[count-1]:\n",
    "                if j in i[1]:\n",
    "                    relevant_retrived+=1\n",
    "    precision.append(relevant_retrived/retrieved)\n",
    "    recall.append(relevant_retrived/len(relevant_documents_id[count-1]))\n",
    "    count+=1\n",
    "plot_precision_recall(precision,recall,\"complete\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Queries on Agglomerative Clustering with Average linkage\n",
    "#### Precison and Recall for each query is calculated and the results are compared with the Precision and Recall values of Task 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdoAAAFSCAYAAABVBx4aAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAblUlEQVR4nO3de7xVdZ3/8dcHUVREEGWUcoisn9VPnSbm2G/GmqzUbipNaJbaRWxASszJvGT9zMvkBaSUwNFBc3IMLSWt/M14mRxRx7EMTcsyrfAGNopKggJi8Pn9sRe4OZzDPsj+nn3O9vV8PM7DxVprr/Xe+8h+8117rbUjM5EkSWUMaHUASZLamUUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq3aQkTsFRFzI+LWiPjPiOjo5f3vFBGnd7PsAxHxkSbv7/0RcW9EnNDM7dZtv9vMEfGliHj9K9zu6IiYs2npmqev5VF7GtjqANKmiojhwIXABzLzDxExFHhDL+4/gCcz89SulmfmDQV2Ow4Yn5k/7+kDImJAZq7uybobypyZ5/R0n81Qvb5kDy/635jnKfUGR7RqB/sDP8jMPwBk5nOZeU9EbBsRP6pGud+NiC0i4t3VvGuqEeHfRcSNEXFXRGxfjXBui4irI+LnEbEvQEScUI2Y50XEPtW8b0fEBcBNQMeakVFE/EtE3F5tZ3REHBERk6tlX4iIOyPijogYU827JyK+HhE/jYiTOj+5iPhYtewn1Uj2vcCHgVkRMbZuve6yz42Ic4Ebo2ZGRNwSET+OiJ2rdcZX+7glIvZbkzkihlePnxsRP6p73rtHxGYR8Z3q9f23iNiuynBrRFwREb+ssnb2Z6/g9d2h7nmOql6/66vf6xF1z/0q4KSI+GT1XO6OiMOrx51WbfOG6nc+tNrkiAZ5pU2Tmf74069/gJOAo7qYfzwwqZo+BfgU8G7gP6p5E4Frq+njgCOB0cDDwCBqb+4/qZZvXf13B2BuNf1tYGI1PRqYA2wO/DcQ1fwBwBHAZGAn4LZq3ui6HPOB1wGbAfd3eg6bAfcBWwLbAvPq9r17p3W7yz4XeF81fQDwtWr6/wAzgRHAncAWXWTeB5i6Zn79voGDgXOqeZ8EvlpleJDa0bK3ANf0MOMGX99O25hZ93y+W2UdXb2Ogzptb0vg7mr6NOCsanoS8MVGef3xpxk/jmjVDp4AXtvF/DcCP6umfwb8r2r6F3WPq5/erpq+PzNfzMynqRUdwOERcRu1Mn1N3T5+Ur/DzHwJmAFcHhHTga3rFo8G7svM1Zn5CDCsmr84Mx/NzFXAik7PYQTwWGauyMwlwEsRsaGPfLrKXp/zfwMfjoi5wNQqwy7Uymhl9RzqD7veCjwfEbOp/WOkXnev7/2Z+SfgcV5+TRtl7NHrW7ffu+v2u8a9mfliNb1f9RxvrMtFp8e9sYd5pU1i0aod/Bu18hgJUB0yHgP8Dnh7tc6ewG+r6frP+uqno/rvbtVh5uHAqmre8cB7gI/VrQewzmeBEbEZMCczPwE8Re2z1DUeAf4yIgZExGjgj11k6GwR8LqI2DIitqU26vzTBtbvKnt9zt8AV2XmuzNzb2A88HtgTERsXj2H+veFzTPzjMw8HHh/RIyqW9aT17f+tdpQxh69vnX7fVs1/VfdrHsqcCDwQWB53fw1j+ug9rx7klfaJJ4MpX4vM5+NiM8CV0ZEUHvzPgG4GJgdER+nVnpTgL16sMkFwJXURnprPjO9BfgvaiOs5zfw2CHAD6uySuDjwPuqnP8TET8E7qiWfb4Hz21VRJxDbWQJ8H9fQfZ61wHvjYhbqgyzM/NbEXEJcEdEPA+cXbf+nhFxJrX3ioer7a/xA2BcNRJ9ATic2uHtRjbl9YXaSPzKiDgeWAa81MU636f2mt0DPFs3/88j4iZqhfpRXj6qIBWz5nMkSdROKAKmZebBLY6y0fpz9o0REQPXjOoj4rvAeZn50x487jRqn3H/v8IRpXV46FhSf/O6qJ3V/RPgjz0pWamVHNFKklSQI1pJkgqyaCVJKsiilSSpoD51ec8OO+yQo0ePbnUMSZI2yt133/10Zo7oalmfKtrRo0czb968VseQJGmjRMSj3S3z0LEkSQX1qRGtJKn1XnrpJRYsWMCKFZ1vva16I0eOZNiwxjcXs2glSetYsGABQ4YMYfTo0dTuaqrOli9fzsKFC3tUtB46liStY8WKFWy//faW7AZsueWWvPRSV7fZXp8jWknSetaUbEfHK3t8d+e1Ll++nC984Qs88cQTLF68mMGDB3PHHXdw+OGHs8suu3DIIYcwc+ZMJk+ezEc+8hHe9a53sWTJEg477DD2228/3v72tzNmzBgAZs6cycCB69bY6tWrGTCg8Rjy4IMPZs6cOWv/fMIJJ3Duuef2+PltzD9CLFpJUq/ZaqutuOiii5g7dy73338/BxxwADNnzmTatGkAPPLII2vX3WeffZg2bRqZydixY9l3330ZNWoUF1100XrbPeKIIxg9ejR77LEHy5cv595772XJkiVccMEFnHnmmSxdupSBAwfy5je/mfHjxwMwf/58zjrrLM477zwefvhhAN75zncybtw45s2bx7Rp03juuec4/fTTedOb3sRPf/pTbrjhho1+zh46liS11I9//GMmTZrE1Vdf3eXyiGCXXXZh0aJFPPbYY0yaNIkzzzxzvfUmTJjAQQcdBMDAgQNZuHAhP//5zwE45JBDmDJlCjfddBMAv/vd7zj77LO54IILGDJkyNptDBkyhOOOO45DDz2UW2+9lUsuuYQpU6Zwyimn9PhQcWeOaCVJLbXvvvt2OaJdIzN5+OGHGTFiRLcjWoChQ4cCcNVVV/GjH/2I008/nWXLlgEwePDgtdsC2H777Vm2bBmLFy9mp512WruNNettvvnmvPjii2QmEbFJn1dbtJKkPunmm2/m2GOPZcmSJRxzzDE9LruRI0cydepU7rrrLvbee+8u19luu+2YMWMGRx99NOedd16325owYQInnXQSu+66K9tss80reh5N+5q8iNgF+AowtP6LpyNid+Dk6o9nZ+b93W2jo6MjvTOUJLXWAw88wFve8pZWx+gTnn32Wc4//3yeeeYZ9tlnH8aNG7d2Wf3rFBF3Z2aXp441bUSbmfOBz0TEnE6LjgWOBhKYChzVrH1KklTS8OHDOeOMMzZpG71x6HhoZv4RICKGNFpZkqR20htF+1xEDKU2ol3aeWFETAQmAowaNaoX4vRfHbN6dkHbvIkefpekvqJpl/dExPYRcRHwtog4OSIurxZNB74JzARmdH5cZs7KzI7M7BgxostvGJIkqd9q5me0zwCTuph/P/DpZu1HktR7enokrbNmH1nr6R2f+iIv75Ek9Zpf/epXzJ49myeffJJRo0YxZswYDjzwQMaPH8+MGTOYPn06ixYtYunSpZx//vkceuih7LXXXuy55548+OCDzJ8/nwEDBjBt2jS+973vcfPNN7PNNtswYMAATjnlFE499VQykyFDhvC1r32t1U8X8M5QkqReNGjQIFasWMGOO+7Io48+yrXXXrv29oiPP/44t912G8OGDWOLLbbggQceYPXq1Zx44onst99+rFq1iq222oo777yTp556iquuuopZs2Zx6KGHAnDFFVewfPlytttuO+bPn8/KlStb/GxrHNFKknrN9OnTOeGEE8hMTj31VDbbbDMuu+wyxo0bx+rVq9ltt9047bTT1q6/9dZbM3DgQJ555hnuu+8+Lr30UsaPH7/2jk/w8g3+V69ezf7778/YsWN7+2ltkEUrSeo173nPe5gyZQo77rgjAAcddBCTJ0/moYceYuDAgQwYMIDjjjuO5cuX8+Uvf3nt44YNG8ayZcuYNm0aDz30EFC7f/FRRx3FoEGDGD58OJ/4xCeYPHkyt99+OytXrmT69OkteY6dNe3OUM3gnaE2zMt7JPWG/nJnqHvuuYfvf//7PPHEE5x88snsuuuuvbr/Xr8zlCRJvWnMmDFrv5u2L/NkKEnSepYvX05fOuLZ16xatarH6zqilSStY+TIkSxcuPAVf//qq8Xw4cN7tJ5FK0lax7Bhwxg2bFirY7QNDx1LklSQRStJUkEWrSRJBVm0kiQVZNFKklSQRStJUkEWrSRJBVm0kiQVZNFKklSQRStJUkEWrSRJBVm0kiQVZNFKklSQRStJUkEWrSRJBVm0kiQVZNFKklSQRStJUkEWrSRJBVm0kiQVZNFKklSQRStJUkEWrSRJBVm0kiQVZNFKklSQRStJUkEWrSRJBVm0kiQVZNFKklSQRStJUkEWrSRJBVm0kiQVZNFKklSQRStJUkEWrSRJBVm0kiQVZNFKklSQRStJUkEWrSRJBVm0kiQVNLBZG4qIwcA/ASuBuZk5u5r/IeBIYDVwSWbe1Kx9SpLU1zVzRDsOmJOZE4CxdfP3Ar4CHAvs08T9SZLU5zVtRAvsDPyyml5VN/9a4NvUSn1C5wdFxERgIsCoUaOaGEeSpNZr5oh2AbWy7bzdk4G9gb8FvtT5QZk5KzM7MrNjxIgRTYwjSVLrNXNEew0wMyL2B66LiMsz85PV/IuBAG5o4v4kSerzmla0mfkCML5u1uxq/hXAFc3ajyRJ/YmX90iSVJBFK0lSQRatJEkFWbSSJBVk0UqSVJBFK0lSQRatJEkFWbSSJBVk0UqSVJBFK0lSQRatJEkFWbSSJBVk0UqSVJBFK0lSQRatJEkFWbSSJBVk0UqSVJBFK0lSQRatJEkFWbSSJBVk0UqSVJBFK0lSQRatJEkFWbSSJBVk0UqSVJBFK0lSQRatJEkFWbSSJBVk0UqSVJBFK0lSQRatJEkFWbSSJBVk0UqSVJBFK0lSQRatJEkFWbSSJBVk0UqSVJBFK0lSQRatJEkFWbSSJBVk0UqSVJBFK0lSQRatJEkFWbSSJBXUo6KNiG0iwlKWJGkjbbA8I+L0iNgeuBK4tHciSZLUPhqNUocBHwT+EVhUPo4kSe2lUdE+CuwB/Ax4unwcSZLaS6OivRn4bWYmcHcv5JEkqa00KtrPA6+vpvff0IoRMTgiLouIiyPi8Lr5IyPimxExIyLesYl5JUnqVxoV7RLgxWp66wbrjgPmZOYEYGzd/OOBpcBqYMErCSlJUn/VqGjvBPaNiBuA2xusuzPweDW9qm7+bsC/AqcBp3R+UERMjIh5ETFv0SLPt5IktZeB3S2IiABeysx39XBbC6iV7b2sW+ALgMXA88CWnR+UmbOAWQAdHR3Zw31JktQvdDuirU6A2qP6jHXriGh06Pga4KCIuBC4LiIur+Z/A5hKrUwvbEZoSZL6i25HtJXXA2fW/fnI7lbMzBeA8XWzZlfzfw0c8QrzSZLUrzUq2onA+6rpGwtnkSSp7TQ6GeoiYHj188/l40iS1F4ajWiXZOblABHx1l7II0lSW2lUtIsj4lvV9KOlw0iS1G4aFe2ZvHyjihcKZ5Ekqe00+oz23MxcmplLgbN6I5AkSe2kUdHWXzu7TckgkiS1o0aHju+IiGuo3af4ul7II0lSW2lUtN/NzMsj4o34fbSSJG20RoeOz4qI1wL/CMzshTySJLWVRkW7LfBh4GzgifJxJElqL42Kdi6wdWb+Avht+TiSJLWXDX5Gm5lX1k1fXD6OJEntpdGIVpIkbYJGZx0TETsCgwAy87HiiSRJaiMbLNqIuBR4BvgTkMCXeyOUJEntotGI9v7M/EavJJEkqQ01KtoPR8QbqL5QIDNPLB9JkqT20ahoP0XtkDFAFM4iSVLbaVS0fwF8ktq9jq/E76SVJGmjNLq858DMPCQzPw58oDcCSZLUThqNaLeKiFHV9ODSYSRJajeNivY04Jhq+oyyUSRJaj/dFm1EvAMYSu1+xwC7Ar/rhUySJLWNDY1oh1c/8PKZx5IkaSN0ezJUZl4H/D4zLwNGA8/3VihJktpFo7OOD4iIvwb+B9i3F/JIktRWGhXt66mdDHUN8GL5OJIktZdGZx0fDQzKzKcj4pLeCCRJUjvZ0FnHU6mNaDMigtoJUYf0VjBJktpBt0XrFwhIkrTpNvgZbUScXjf9lfJxJElqL41Ohtqubnr7kkEkSWpHjU6GeiwiLqP27T2/7oU8kiS1lUZF+w3gtcBiYGX5OJIktZdGh47PB07LzOeBGb2QR5KkttKoaFfx8pe9P1c4iyRJbadR0b4IvCUiJrPuiVGSJKkHGn1G+y3gDdQK+cLycSRJai+NivYjmTm1V5JIktSGGhXtpyJiH2AJkJnpLRglSdoIGyzazNw9IiIz/eJ3SZJegQ0WbUQcB3RExObAXZl5bu/EkiSpPTQ663izzDwsMz8K7NgbgSRJaieNPqPdLSIOq9YbGREfysx/74VckiS1hUZFewuweTV9E7BD2TiSJLWXRidDXdZbQSRJakeNPqOVJEmbwKKVJKmgphVtRAyOiMsi4uKIOLzTsj0i4qmI2KZZ+5MkqT9o5oh2HDAnMycAY9fMrK7B/Xvg+ibuS5KkfqGZRbsz8Hg1vapu/vHUvsu2y7tLRcTEiJgXEfMWLVrUxDiSJLVeM4t2AbWy7bzdvwQmA28Hjur8oMyclZkdmdkxYsSIJsaRJKn1mlm01wAHRcSFwHURcTlAZn4sM/8BuAv45ybuT5KkPq/RDSt6LDNfAMbXzZrdafkRzdqXJEn9hZf3SJJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJU0MBmbSgiBgP/BKwE5mbm7Gr+l4A3ADsBn83MBc3apyRJfV0zR7TjgDmZOQEYu2ZmZp5TzbsUeE8T9ydJUp/XzKLdGXi8ml5VvyAitgEOAX7Q+UERMTEi5kXEvEWLFjUxjiRJrdfMol1ArWzX2W5EbAtcCJyYmUs7PygzZ2VmR2Z2jBgxoolxJElqvWYW7TXAQRFxIXBdRFxezf82MAz4SkS8t4n7kySpz2vayVCZ+QIwvm7W7Gr+uGbtQ5Kk/sbLeyRJKsiilSSpIItWkqSCLFpJkgqyaCVJKqhpZx1Laqyjo+frzptXLoek3uOIVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgga2OsCrXUfHRqw8sVgMSVIhjmglSSrIopUkqSCLVpKkgvyMVpIK2ZhzMObNK5dDreWIVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrI62gl9Us9vkZ1Ys8vZp030YtZ1XwWrSSpz+vpP6z64o0/LFr1SH/4n7w/ZJT6Iv/ulNXWRev/PJKkVmvrou2pjll+hvNq4u+7Mf+RKjWPZx1LklRQ04o2IgZHxGURcXFEHF43f/eImF397N6s/UmS1B80c0Q7DpiTmROAsXXzjwWOBj4HHNPE/UmS1OdFZjZnQxEnA9dn5r0RcUVmHlbNvyozD6mm186ve9xEYGL1xzcBDzYlUPd2AJ4uvI9m6A85zdg8/SGnGZunP+Q048Z5XWaO6GpBM0+GWgDsDNzLuiPl5yJiKJDA0s4PysxZwKwm5tigiJiXmRvxdcyt0R9ymrF5+kNOMzZPf8hpxuZpZtFeA8yMiP2B6yLi8sz8JDAd+CYQwNQm7k+SpD6vaUWbmS8A4+tmza7m3w98uln7kSSpP3k1Xt7Ta4epN1F/yGnG5ukPOc3YPP0hpxmbpGknQ0mSpPW9Gke0kiT1mra/BWNEbAF8HVgFbAvMBA4CPgqMycznWxgP6Dbj54HNgEWZeVwL4wHdZvwY8GfAVsCnMnNl6xLWdJUzM++JiC8BHZl5cEsD0u1reRFwD/BoZp7dwnhAtxl3Bt4HvAScnJkrWpewpoucFwFHVov3BfbPzNKXDG5QN6/lkcCWwFBgfKvfh7rIeAEwoVq8DPhitujwZ0/ewyPiNcC51Tr/kpm3tCJrd9q+aIG/B/49M6+PiIHAlZn50Yh4bauD1ekyI0BEXB0RAzJzdWsjbjDj14EhwDOtDFhZL2eV7w8tzlVvvYzAC8AWwBMtTfayrjJuC9wHLOkLJVvp7u/3FsDVrS7ZSlevJVXOk4E/Bx5oZUDWz3g98FhmfiYiTgTeAfxXH8nW1Xv4Z4BzgF8B3wH6VNG+Gg4d7w78DCAz/wT8qfpl9SVdZoyIvwV+0wdKFrrOuHNEfAcYRRfXSLdI55wDgU9n5mUtTbWu9V5L4P2ZeSTwoYgY3spwlc4ZdwCGZOaJwOKIeG8rw9Xp7u/33wE/bGWwOl39vh+MiOuBvwF+28Jsa3TO+DS1jOcDu1E7mtFXsnX1Hr4z8Hgfea9cz6uhaH8N/BVA9ct5TfXL6kvWywi8k9qbxVdbmKteV6/jgsz8BLW/BH/TynB1OuccCwyu3jDeGhF/3cpwla5eyzWH3RdTO6TYap0zbsnLo+3F1I5g9AXd/f3+OPC9Vgar0znjX1DL+UHgKuCAFmZbo6vXcWpm/gPwGOXv2Lex2Tq/hy8Ado6IPtlpbX/WcUQMonZ8fzW1X9ZFwAjgKOB24KuZ2dLDdV1knEXt84Zrqd1R6wuZubx1CbvNuCe1fIOByZm5rHUJa7r6fWfm5dWyOX3kM9qu/p/cD1gBPJuZX2phPKDbjNsBb6BWsp/NzBdbl7Cmm5x3ACdm5qRWZlujm4x7AyuBnYBjMnNh6xJ2m3FXakcynsrMU/tYtnXew6tVz6F2tOA7mfmfLYjarbYv2nrVZ3WXZ+a9rc7SHTM2T3/Iacbm6Q85zbhp+nK2DXlVFa0kSb2tTx7PliSpXVi0kiQVZNFKklRQX7ueVFIPdHcHrB4+NoADgcWZeXu5lJLAk6GkfikiPgc83OluQ5GZB0fEJOA3wHPAEdT+QX0ntWsNT6B2+cuT1c8vgS9S+77o3wM3AqdSu3byX6uvuZS0CTx0LPVPXd1tqLPjqN1cYhHwtmref2fmWdRGwgCfA5ZTu33mHtTuvbsYuMKSlZrDQ8dS/7Tmbjk31t1N7Mlq2eDqv1sA0zNzMUBEvJvaKLfeAGrXJf5izYyIWAh8JiLe2sduXSn1Sxat1D9dDHw9Ivbn5bvljImI46ndvvNuYAowIyKeBB6hdpi4s5nAWRHxB2r3q76N2ue3w4H/KP0kpFcDP6OV+rn+ercc6dXCopUkqSBPhpIkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgv4/tmtdNyfif4MAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cluster_single = AgglomerativeClustering(n_clusters=30, affinity='euclidean',linkage='average')\n",
    "cluster_single.fit_predict(vectorized_data_docs)\n",
    "clusters_single = cluster_single.labels_.tolist()\n",
    "clusters_single_dict=packing_data_into_clusters(clusters_single)\n",
    "centroid_single=[]\n",
    "precision=[]\n",
    "recall=[]\n",
    "for i in clusters_single_dict:\n",
    "    centroid_single.append(centroid(i[1]))\n",
    "count=1\n",
    "for q in vectorized_data_queries:\n",
    "    cos_single=[]\n",
    "    for i in range(len(centroid_single)):\n",
    "        cos_single.append(cos(centroid_single[i],q))\n",
    "    index=cos_single.index(max(cos_single))\n",
    "    for i in clusters_single_dict:\n",
    "        if i[0] == index:            \n",
    "            retrieved=len(i[1])\n",
    "            relevant_retrived=0\n",
    "            for j in relevant_documents_id[count-1]:\n",
    "                if j in i[1]:\n",
    "                    relevant_retrived+=1\n",
    "    precision.append(relevant_retrived/retrieved)\n",
    "    recall.append(relevant_retrived/len(relevant_documents_id[count-1]))\n",
    "    count+=1\n",
    "plot_precision_recall(precision,recall,\"average\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Queries on Agglomerative Clustering with Ward's Method\n",
    "#### Precison and Recall for each query is calculated and the results are compared with the Precision and Recall values of Task 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAdoAAAFSCAYAAABVBx4aAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAZ/ElEQVR4nO3dfbRcdX3v8fc3hBgaQxA4t4nFmIvlqeITPdpqMaBEKgXTGmh8QBSwCSiPooAsL7XoCghEwBJMbqAWi1EuZKFCb1Ua2yBiFYOEypVqVRCDChEiIEkw5nzvH7MPTCbnnDkx+3fmnMn7tdZZ2fnNnr0/M5Mzn/z2zOyJzESSJJUxrtMBJEnqZhatJEkFWbSSJBVk0UqSVJBFK0lSQRatJEkFWbTaIUTEayJiZUTcFhH/FhG9I7z/qRFxwSCXvTEi3lzz/v48IlZHxNl1brdp+9dGxIERcWhELGy5bKuxThptebTjGd/pAFJpEbE7sBh4Y2b+PCKmAC8awf0H8HBmfnigyzPzywV2Owc4ITPvHiTTuMzsK7Df38m25hlt+aWhOKPVjuBI4AuZ+XOAzHw8M78TEbtGxM3VLPf6iJhQzX5ujoibqhnhX0XEVyLizojYIyJmRMTXIuLGiLg7ImYBRMTZ1Yx5VUQcVo1dGxFXAbcCvRGxvBr/x4i4vdrOjIg4PiJOrS57X0T8R0TcEREHVWPfiYiPR8S3IuLc1hsXEW+pLvtmNZN9PfCXwNKImN20Xn/2G4BzI2LviLi1yn15tc4uEfG56j5ZMdhtG4b9qvtxVUS8rNrOZdV2vxURL63GVkbEpcBXWm7TK6rr3lL9HFr93BIRXwCOH+o+j4gV1WPUP5nYp3pM74mIA4d5G6R6ZKY//nT1D3AucNIA4x8ATq6WzwfeCRwK/Gs1Nh/4fLV8FnAiMAO4H3gOsCfwzery36v+3BNYWS1fC8yvlmcAy4GdgW8AUY2PA44HTgWmAl+rxmY05fgx8EJgJ+DeltuwE3APMBHYFVjVtO8DW9adUW3rOdXfbwD2rZYXA73A6cD7+7fd5rYdWN1fC1v2cyjwdSCAA4BbWrbzcuAz1fJK4PABHpt/Bvat7otvVNvs3+64Yd7nHwOOrq731WrsCOCyTv+b9GfH+vHQsXYEPwP2GWD8D4Grq+VvA38GPAj8Z9P1mpf/oFq+NzOfBp6OiJ2qsWMj4jigD3h+0z6+2bzDzNwUEVcC10XEo8CHmi6eAdyTjUOiD0TEbtX4usz8CUBEbGy5DT3Ag5m5EdgYEZuaZnEDWV1lB9ifxqwXYDKNWeUBwD9UWTe3uW1DuTszE7gvIqZWY++PiMOr5d82rftNtvb7mfkDgIi4q2n8znz2kPFgufrX/zaNx/hRYHU19lPgecO8DVItPHSsHcH/Bf4yIqYBVIeMDwJ+CLyqWueVwH9Xy80nAG9ejurPF1eHmXcH+svoA8DrgLc0rQeNEnh2A41iXp6Z7wAeofFaar8HgJdHxLiImAH8aoAMrdYCL4yIiRGxKzAhM387xPrNeb4PvCszD6Uxm/1n4D5gZpW1//lhsNs2lJdHw37ALyJiD+DIzHwtcCZD3EeVhyNin2j8L+CgQdYdLNcrqj97gR9VywM9jtKIcEarrpeZj0XEe4DPVU/cm4Gzacxml0XEW2mU3sXAa4axyTXA54C9aRyWBvh3Goc1vwn8eojrTga+WJVYAm8FDq9y/iIivgjcUV12+jBu2+aI+BhwWzX0v4aRv9+5wJKImEjjPjmRxn1ybUTcBmwCZm3DbWv2OHALjcPh84B1wCMRsZKBZ7CtzqdxHz8MPFll2bllncFy/XFEvI3Gf1TOBw4eZmapiP7XiSQNQzXTXJiZx3Q4SleLiJ2rw+zjaPwnYm5Wb2Zrc71raTw+95bOKA2Xh44ljUZ/Us2q7wS+MpySlUYrZ7SSJBXkjFaSpIIsWkmSCrJoJUkqaFR9vGfPPffMGTNmdDqGJEnb5K677vplZvYMdNmoKtoZM2awatWqTseQJGmbRMRPBrvMQ8eSJBU0qma0kqTO27RpE2vWrGHjxtZTa6vZtGnT2G233dquZ9FKkrawZs0aJk+ezIwZM6i+dEItNmzYwEMPPTSsovXQsSRpCxs3bmSPPfawZIcwceJENm3aNKx1ndFKkrbSX7K9vb/b9Qd7X+uGDRt43/vex89+9jPWrVvHpEmTuOOOOzj22GPZe++9mTt3LosWLeLUU0/lzW9+MzNnzuSJJ57g7W9/O294wxt41atexUEHNb7QadGiRYwfv2WN9fX1MW5c+znkMcccw/Lly5/5+9lnn82ll1467Nu3Lf8JsWglSSNml112YcmSJaxcuZJ7772Xo446ikWLFrFw4UIAHnjggWfWPeyww1i4cCGZyezZs5k1axbTp09nyZIlW233+OOPZ8aMGbzkJS9hw4YNrF69mieeeIKrrrqKBQsW8OSTTzJ+/Hj2339/TjjhBAB+/OMfc+GFF3L55Zdz//33A3DwwQczZ84cVq1axcKFC3n88ce54IIL2G+//fjWt77Fl7/85W2+zR46liR11IoVKzj55JO58cYbB7w8Ith7771Zu3YtDz74ICeffDILFizYar158+Zx9NFHAzB+/Hgeeugh7r77bgDmzp3LxRdfzK233grAD3/4Qy666CKuuuoqJk+e/Mw2Jk+ezFlnncXb3vY2brvtNq655houvvhizj///GEfKm7ljFaS1FGzZs0acEbbLzO5//776enpGXRGCzBlyhQAbrjhBm6++WYuuOAC1q9fD8CkSZOe2RbAHnvswfr161m3bh1Tp059Zhv96+288848/fTTZCYRsV2vV1u0kqRR6atf/SpnnHEGTzzxBKeddtqwy27atGlccskl3HnnnRxyyCEDrvO85z2PK6+8klNOOYXLL7980G3NmzePc889l3333ZfnPve5v9PtqO1r8iJib+BDwJTmL8WOiAOB86q/XjTUFzL39vamZ4aSpM667777OOCAAzodY1R47LHHuOKKK3j00Uc57LDDmDNnzjOXNd9PEXFXZg741rHaZrSZ+WPg3RGxvOWiM4BTgAQuAU6qa5+SJJW0++6785GPfGS7tjESh46nZOavACJicruVJUnqJiNRtI9HxBQaM9onWy+MiPnAfIDp06fXuuNhf/5r/vA/KLZqfr2HtrflM2qdPKpe931Z9/0IPt51Gm7OsZBxrDzeo+l355JLYKBXFf/oj2rf9bB973vDW6+TGQdT28d7ImKPiFgCvCIizouI66qLPgH8PbAIuLL1epm5NDN7M7O3p2fAbxiSJGnMqvM12keBkwcYvxd4V137kSSNnLkrnp3uTvz68K9Xx8y79SxPZ555DFdc0fo2oNHPj/dIkkbMGWecwUc/+lHOOussXve61zFhwgQmTJjAHXfcwSOPPMJll13GzTffzIoVK+jt7eWggw5i6dKl7LbbPp2O/jvzzFCSpBFz8MEHc/vttzNlyhTuuecebr/9dtavX09fXx+bNm1ixYoVABxxxBGceeaZLF68mGuuuYa5c7c6YDpmWLSSpBFz6KGH8qlPfYp99tmHvr4+HnvsMa6//noWLlzI4Ycf/syZnPrP8jRu3DjGjx/PhAnP6WTs7eKhY0nSiOnp6eEHP/gBCxYsYOPGjfT19fGCF7yABQsWcN999zFr1qwt1j/ppJM455xziJg6yBZHP4tWkjSoG2Y9+6amuj46893vfheA/fffv+26M2fOZObMmcP+eM9o5KFjSZIKsmglSVvo64O+vg3UdS78brR58+Zhr+uhY0nSFq6/fhpHHvkQu+66ieYvzNmOb4rbbr/4xfDWG8mMu++++7DWs2glSVtYvXo3Vq/ebavxTp5y87jjhrfeaPwCOA8dS5JUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFTS+rg1FxCTgk8BvgJWZuawa/wvgRKAPuCYzb61rn5IkjXZ1zmjnAMszcx4wu2n8NcCHgDOAw2rcnyRJo15tM1pgL+C71fLmpvHPA9fSKPV5rVeKiPnAfIDp06fXGEeSpM6rc0a7hkbZtm73POAQ4LXAB1uvlJlLM7M3M3t7enpqjCNJUufVOaO9CVgUEUcCt0TEdZl5XDV+NRDAl2vcnyRJo15tRZuZTwEnNA0tq8Y/C3y2rv1IkjSW+PEeSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIKGl/XhiJiEvBJ4DfAysxcVo1PA84DArg+M++oa5+SJI12dc5o5wDLM3MeMLtp/APAk0AfsKbG/UmSNOoNWbQRcXD15/kRMafNtvYCflotb24afzHwT8DfAecPsI/5EbEqIlatXbt2uLklSRoT2s1oj4qIPwUeBma1WXcNjbJt3e4aYB3wa2Bi65Uyc2lm9mZmb09Pz/BSS5I0RrQr2v8JnAbcBDzdZt2bgKMjYjFwS0RcV41fBlwCLAUWb0dWSZLGnHZvhjoF2CUzfxkR/zDUipn5FHBC09Cyavx7wPHbE1KSpLGq3Yz2DOCiavndhbNIktR12hXtZOAH1fKmwlkkSeo67Yo2gWkRcRQwdQTySJLUVdoV7eXAPcB04NTycSRJ6i7t3gz11sy8ZESSSJLUhdoV7eyIeBHwOEBmnlM+kiRJ3aNd0R5L43VaaJyrWJIkbYN2RftS4Dga5yn+HPCT4okkSeoi7d4M9abMnJuZbwXeOBKBJEnqJu1mtLtExPRqeVLpMJIkdZt2Rft3NM51DPCRslEkSeo+7Yr2gMw8GyAijgB+WD6SJEndo91rtIc0Lb+2ZBBJkrpRuxltT0QcRnUqxhHII0lSV2k3oz0d2BfYDzizfBxJkrrLcA4dL6Xxedp3lI8jSVJ3aXfo+GDgSeCrwGvKx5Ekqbu0m9G+gMbHe26lfSlLkqQW7crzXcBOmbkxIi4ciUCSJHWTIYs2MzcBm6rlX4xIIkmSuki7Q8eSJGk7DDqjjYhL2fIr8tLvo5UkadsMdeh4Ucvfc8C1JEnSoIYq2lNpmdECzmglSdoGzmglSSpo0KLNzJ9ExJ8AxwG/Vw2fOCKpJEnqEu3edTwP+BWN76W9v3gaSZK6TLuifRiYCPQBv18+jiRJ3aXdmaGWAU/TeBPUivJxJEnqLu1mtHtn5o8y83QahStJkrbBcL4mr99rSwaRJKkbtTt03BMRh9H4aM+0EcgjSVJXaTejPR3YF9gPOLN8HEmSuku7on0TjddpFwPvGIE8kiR1lXZF+2rgyWp5RtkokiR1n3ZF+1uAiJgCTC0fR5Kk7tKuaK8F/hBYAlxWPI0kSV1mqO+jDWBmZr5zBPNIktRVhvpSgYyIV0bE24DHq7F/GbFkkiR1gXafo10BTAB68GvyJEnaZkMWbWZ+eqSCSJLUjdq9GUqSJG0Hi1aSpIIsWkmSCrJoJUkqyKKVJKkgi1aSpIIsWkmSCrJoJUkqqLaijYhJEfHpiLg6Io5tuewlEfFIRDy3rv1JkjQW1DmjnQMsz8x5wOz+wYjYGfgb4Es17kuSpDGhzqLdC/hptby5afwDwJUMcq7kiJgfEasiYtXatWtrjCNJUufVWbRraJRt63ZfDpwKvAo4qfVKmbk0M3szs7enp6fGOJIkdV6dRXsTcHRELAZuiYjrADLzLZl5JnAn8L9r3J8kSaNeu6/JG7bMfAo4oWloWcvlx9e1L0mSxgo/3iNJUkEWrSRJBVm0kiQVZNFKklSQRStJUkEWrSRJBVm0kiQVZNFKklSQRStJUkEWrSRJBVm0kiQVZNFKklSQRStJUkEWrSRJBdX2NXkqr3dp77DWWzV/VeEkkqThckYrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQVZtJIkFWTRSpJUkEUrSVJBFq0kSQWNr2tDETEJ+CTwG2BlZi6rxj8IvAiYCrwnM9fUtU9Jkka7Ome0c4DlmTkPmN0/mJkfq8Y+Bbyuxv1JkjTq1Vm0ewE/rZY3N18QEc8F5gJfaL1SRMyPiFURsWrt2rU1xpEkqfPqLNo1NMp2i+1GxK7AYuCczHyy9UqZuTQzezOzt6enp8Y4kiR1Xp1FexNwdEQsBm6JiOuq8WuB3YAPRcTra9yfJEmjXm1vhsrMp4ATmoaWVeNz6tqHJEljjR/vkSSpIItWkqSCLFpJkgqyaCVJKsiilSSpIItWkqSCLFpJkgqyaCVJKsiilSSpIItWkqSCLFpJkgqyaCVJKsiilSSpIItWkqSCLFpJkgqyaCVJKsiilSSpIItWkqSCLFpJkgqyaCVJKsiilSSpIItWkqSCLFpJkgqyaCVJKsiilSSpIItWkqSCLFpJkgqyaCVJKsiilSSpIItWkqSCLFpJkgqyaCVJKsiilSSpIItWkqSCLFpJkgqyaCVJKsiilSSpIItWkqSCLFpJkgqyaCVJKsiilSSpIItWkqSCLFpJkgqyaCVJKsiilSSpIItWkqSCLFpJkgqyaCVJKsiilSSpoNqKNiImRcSnI+LqiDi2afzAiFhW/RxY1/4kSRoL6pzRzgGWZ+Y8YHbT+BnAKcB7gdNq3J8kSaNeZGY9G4o4D/hSZq6OiM9m5tur8Rsyc261/Mx40/XmA/Orv+4HfL+WQIPbE/hl4X3UYSzkNGN9xkJOM9ZnLOQ047Z5YWb2DHTB+Bp3sgbYC1jNljPlxyNiCpDAk61XysylwNIacwwpIlZlZu9I7e93NRZymrE+YyGnGeszFnKasT51Fu1NwKKIOBK4JSKuy8zjgE8Afw8EcEmN+5MkadSrrWgz8ynghKahZdX4vcC76tqPJEljyY748Z4RO0y9ncZCTjPWZyzkNGN9xkJOM9aktjdDSZKkre2IM1pJkkZMnW+GGpUiYgLwcWAzsCuwCDga+GvgoMz8dQfjAYNmPB3YCVibmWd1MB4waMa3AP8D2AV4Z2b+pnMJGwbKmZnfiYgPAr2ZeUxHAzLofbkE+A7wk8y8qIPxgEEz7gUcDmwCzsvMjZ1L2DBAziXAidXFs4AjM7P0RwaHNMh9eSIwEZgCnNDp56EBMl4FzKsuXg+8Pzt0+HM4z+ER8Xzg0mqdf8zMf+9E1sF0fdECfwP8S2Z+KSLGA5/LzL+OiD/odLAmA2YEiIgbI2JcZvZ1NuKQGT8OTAYe7WTAylY5q3w/73CuZltlBJ4CJgA/62iyZw2UcVfgHuCJ0VCylcF+vycAN3a6ZCsD3ZdUOc8DXgDc18mAbJ3xS8CDmfnuiDgH+DPg66Mk20DP4e8GPgb8P+AzwKgq2h3h0PGBwLcBMvO3wG+rB2s0GTBjRLwW+K9RULIwcMa9IuIzwHQG+Ix0h7TmHA+8KzM/3dFUW9rqvgT+PDNPBP4iInbvZLhKa8Y9gcmZeQ6wLiJe38lwTQb7/f4r4IudDNZkoMf7+xHxJeDVwH93MFu/1oy/pJHxCuDFNI5mjJZsAz2H7wX8dJQ8V25lRyja7wF/DFA9OM+vHqzRZKuMwME0niz+toO5mg10P67JzHfQ+CV4dSfDNWnNORuYVD1hvCwi/rST4SoD3Zf9h93X0Tik2GmtGSfy7Gx7HY0jGKPBYL/fbwX+TyeDNWnN+FIaOY8AbgCO6mC2fgPdj5dk5pnAg5Q/Y9+2Zmt9Dl8D7BURo7LTuv5dxxHxHBrH9/toPFhLgB7gJOB24G8zs6OH6wbIuJTG6w2fp3FGrfdl5obOJRw04ytp5JsEnJqZ6zuXsGGgxzszr6suWz5KXqMd6N/kG4CNwGOZ+cEOxgMGzfg84EU0SvY9mfl05xI2DJLzDuCczDy5k9n6DZLxEOA3wFTgtMx8qHMJB824L40jGY9k5odHWbYtnsOrVT9G42jBZzLz3zoQdVBdX7TNqtfqrsvM1Z3OMhgz1mcs5DRjfcZCTjNun9GcbSg7VNFKkjTSRuXxbEmSuoVFK0lSQRatJEkFjbbPk0oahsHOgDXM6wbwJmBdZt5eLqUk8M1Q0pgUEe8F7m8521Bk5jERcTLwX8DjwPE0/kP9HzQ+a3g2jY+/PFz9fBd4P43vi/4R8BXgwzQ+O/lP1ddcStoOHjqWxqaBzjbU6iwaJ5dYC7yiGvtGZl5IYyYM8F5gA43TZ76Exrl31wGftWSlenjoWBqb+s+W85Wms4k9XF02qfpzAvCJzFwHEBGH0pjlNhtH43OJ/9k/EBEPAe+OiJeNslNXSmOSRSuNTVcDH4+II3n2bDkHRcQHaJy+8y7gYuDKiHgYeIDGYeJWi4ALI+LnNM5X/TUar9/uDvxr6Rsh7Qh8jVYa48bq2XKkHYVFK0lSQb4ZSpKkgixaSZIKsmglSSrIopUkqSCLVpKkgixaSZIK+v+XLLgkt7ZMlAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "cluster_single = AgglomerativeClustering(n_clusters=37, affinity='euclidean',linkage='ward')\n",
    "cluster_single.fit_predict(vectorized_data_docs)\n",
    "clusters_single = cluster_single.labels_.tolist()\n",
    "clusters_single_dict=packing_data_into_clusters(clusters_single)\n",
    "centroid_single=[]\n",
    "precision=[]\n",
    "recall=[]\n",
    "for i in clusters_single_dict:\n",
    "    centroid_single.append(centroid(i[1]))\n",
    "count=1\n",
    "for q in vectorized_data_queries:\n",
    "    cos_single=[]\n",
    "    for i in range(len(centroid_single)):\n",
    "        cos_single.append(cos(centroid_single[i],q))\n",
    "    index=cos_single.index(max(cos_single))\n",
    "    for i in clusters_single_dict:\n",
    "        if i[0] == index:            \n",
    "            retrieved=len(i[1])\n",
    "            relevant_retrived=0\n",
    "            for j in relevant_documents_id[count-1]:\n",
    "                if j in i[1]:\n",
    "                    relevant_retrived+=1\n",
    "    precision.append(relevant_retrived/retrieved)\n",
    "    recall.append(relevant_retrived/len(relevant_documents_id[count-1]))\n",
    "    count+=1\n",
    "plot_precision_recall(precision,recall,\"ward\")"
   ]
  }
 ],
 "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}