depression_vs_non_depression_detection.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "depression_vs_non_depression_detection",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "H0nOKFsQXRfU"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import tensorflow as tf\n",
        "from sklearn.model_selection import train_test_split\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "from keras.layers.convolutional import Conv1D\n",
        "from keras.layers.convolutional import MaxPooling1D\n",
        "from keras.layers.embeddings import Embedding\n",
        "from keras.layers import Flatten\n",
        "\n",
        "from keras.preprocessing.text import Tokenizer\n",
        "from keras.preprocessing.sequence import pad_sequences\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sb\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn.metrics import classification_report\n",
        "from sklearn.metrics import plot_confusion_matrix\n",
        "\n",
        "from keras.constraints import maxnorm\n",
        "from keras.layers import Dropout\n",
        "from tensorflow.keras.utils import plot_model\n",
        "import time\n",
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "465blQZ_X718"
      },
      "source": [
        "path = '/content/drive/MyDrive/final_tidy_reddits_lemmetized.json'\n",
        "df = pd.read_json(path, orient='table')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VxnpFu0mpFDY",
        "outputId": "68c7c11e-9a46-4256-b4f3-109a8ee5bccf"
      },
      "source": [
        "df.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(301746, 4)"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mT3B6wV9CrIb"
      },
      "source": [
        "pd.options.display.max_colwidth = None\n",
        "pd.options.display.max_rows = None"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "fg0u1SJ4rIul",
        "outputId": "9dad3fb3-d55d-4547-efa1-5009619e520f"
      },
      "source": [
        "df.sample(7)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>label</th>\n",
              "      <th>text</th>\n",
              "      <th>lemmatized</th>\n",
              "      <th>tokens_back_to_text</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>129582</th>\n",
              "      <td>depression</td>\n",
              "      <td>[hi, guys, maybe, help, think, depressed, life, never, thought, depression, cause, always, thought, feelings, normal, pretty, sure, used, cry, every, single, day, every, little, thing, would, make, cry, fourth, grade, still, know, like, extremely, sensitive, never, cared, school, made, never, homework, pretty, much, life, care, could, drive, social, anxiety, think, started, parents, telling, useless, would, never, amount, anything, never, thought, suicide, always, negative, thoughts, everything, never, remember, ever, happy, im, kind, still, like, except, way, less, severe, anyone, experienced, things, thank, advance, think, depressed, since, child]</td>\n",
              "      <td>[hi, guy, maybe, help, think, depressed, life, never, thought, depression, cause, always, thought, feeling, normal, pretty, sure, use, cry, every, single, day, every, little, thing, would, make, cry, fourth, grade, still, know, like, extremely, sensitive, never, care, school, make, never, homework, pretty, much, life, care, could, drive, social, anxiety, think, start, parent, tell, useless, would, never, amount, anything, never, thought, suicide, always, negative, thought, everything, never, remember, ever, happy, im, kind, still, like, except, way, less, severe, anyone, experienced, thing, thank, advance, think, depressed, since, child]</td>\n",
              "      <td>hi guy maybe help think depressed life never thought depression cause always thought feeling normal pretty sure use cry every single day every little thing would make cry fourth grade still know like extremely sensitive never care school make never homework pretty much life care could drive social anxiety think start parent tell useless would never amount anything never thought suicide always negative thought everything never remember ever happy im kind still like except way less severe anyone experienced thing thank advance think depressed since child</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>199105</th>\n",
              "      <td>depression</td>\n",
              "      <td>[miserable, life, envy, people, lives]</td>\n",
              "      <td>[miserable, life, envy, people, life]</td>\n",
              "      <td>miserable life envy people life</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>299657</th>\n",
              "      <td>depression</td>\n",
              "      <td>[people, may, see, perfectly, normal, person, see, deformed, monster, see, matter, wear, body, fear, smell, im, hideous, expect, people, run, ive, treated, like, freak, nature, vultures, picked, away, flaws, im, disgusted, entirety, body, im, sure, society, become, critical, peoples, appearances, im, treated, like, never, made, cut, demons, treating, apart, anymore, years, self, confident, ive, made, feel, worthless, unlovable, unattractive, years, finding, difficult, function, society, im, constantly, shot, treat, like, im, shadow, ive, finally, fallen, rock, bottom, blazing, molten, core, get, home, everyday, scream, mirror, sit, silence, pass, exhaustion, repeat, day, going, long, im, longer, alive, im, zombie, shuffling, way, distorted, ...]</td>\n",
              "      <td>[people, may, see, perfectly, normal, person, see, deform, monster, see, matter, wear, body, fear, smell, im, hideous, expect, people, run, ive, treat, like, freak, nature, vulture, picked, away, flaw, im, disgust, entirety, body, im, sure, society, become, critical, people, appearance, im, treat, like, never, make, cut, demon, treat, apart, anymore, year, self, confident, ive, make, feel, worthless, unlovable, unattractive, year, find, difficult, function, society, im, constantly, shot, treat, like, im, shadow, ive, finally, fall, rock, bottom, blazing, molten, core, get, home, everyday, scream, mirror, sit, silence, pas, exhaustion, repeat, day, go, long, im, longer, alive, im, zombie, shuffle, way, distort, ...]</td>\n",
              "      <td>people may see perfectly normal person see deform monster see matter wear body fear smell im hideous expect people run ive treat like freak nature vulture picked away flaw im disgust entirety body im sure society become critical people appearance im treat like never make cut demon treat apart anymore year self confident ive make feel worthless unlovable unattractive year find difficult function society im constantly shot treat like im shadow ive finally fall rock bottom blazing molten core get home everyday scream mirror sit silence pas exhaustion repeat day go long im longer alive im zombie shuffle way distort reality birthday week dont want see time finally slip void finally time let go</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>462176</th>\n",
              "      <td>bipolar</td>\n",
              "      <td>[im, single, years, certain, point, life, dont, relationship, nothing, else, matters, almost, reached, point, realize, want, find, one, continue, working, work, find, better, job, prospects, constantly, improving, seems, moot, hard, go, unloved, dramatic, im, feeling, many, memes, subreddit, doesnt, really, feel, genuine, lonely]</td>\n",
              "      <td>[im, single, year, certain, point, life, dont, relationship, nothing, else, matter, almost, reach, point, realize, want, find, one, continue, work, work, find, well, job, prospect, constantly, improve, seem, moot, hard, go, unloved, dramatic, im, feel, many, meme, subreddit, doesnt, really, feel, genuine, lonely]</td>\n",
              "      <td>im single year certain point life dont relationship nothing else matter almost reach point realize want find one continue work work find well job prospect constantly improve seem moot hard go unloved dramatic im feel many meme subreddit doesnt really feel genuine lonely</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>196476</th>\n",
              "      <td>depression</td>\n",
              "      <td>[online, relationship, love, life, problem, able, tell, upset, want, go, numb, upsets, talk, asks, want, never, say, going, upset, met, world, warcraft, play, tons, games, together, lot, fun, miss, family, though, ocd, makes, even, harder, tell, want, family, feel, like, upset, wanted, alone, would, ruin, know, could, right, hard, short, ca, tell, people, really, feel, due, scared, upset, slightest, ways, new, could, use, advice]</td>\n",
              "      <td>[online, relationship, love, life, problem, able, tell, upset, want, go, numb, upset, talk, asks, want, never, say, go, upset, met, world, warcraft, play, ton, game, together, lot, fun, miss, family, though, ocd, make, even, harder, tell, want, family, feel, like, upset, want, alone, would, ruin, know, could, right, hard, short, ca, tell, people, really, feel, due, scar, upset, slightest, way, new, could, use, advice]</td>\n",
              "      <td>online relationship love life problem able tell upset want go numb upset talk asks want never say go upset met world warcraft play ton game together lot fun miss family though ocd make even harder tell want family feel like upset want alone would ruin know could right hard short ca tell people really feel due scar upset slightest way new could use advice</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>125810</th>\n",
              "      <td>depression</td>\n",
              "      <td>[even, sure, counts, depression, year, old, girl, struggled, moderate, anxiety, early, years, old, last, year, still, struggle, much, since, anxiety, subsided, inconsistent, bouts, depression, tiniest, things, trigger, depression, extremely, intense, comes, along, depression, get, extremely, angry, lash, isolate, tend, issues, socially, periods, really, feeling, constant, tiredness, get, tired, easily, wake, start, day, feel, need, go, back, bed, within, hours, feel, destructive, towards, others, think, heavily, suicide, breakdowns, screaming, crying, hair, pulling, kind, make, feel, like, dying, last, days, week, two, never, amount, time, get, happy, periods, thing, happy, periods, really, sure, actually, happy, get, guilty, happy, feel, weird, usually, manifests, ...]</td>\n",
              "      <td>[even, sure, count, depression, year, old, girl, struggle, moderate, anxiety, early, year, old, last, year, still, struggle, much, since, anxiety, subside, inconsistent, bout, depression, tiniest, thing, trigger, depression, extremely, intense, come, along, depression, get, extremely, angry, lash, isolate, tend, issue, socially, period, really, feel, constant, tiredness, get, tire, easily, wake, start, day, feel, need, go, back, bed, within, hour, feel, destructive, towards, others, think, heavily, suicide, breakdown, scream, cry, hair, pull, kind, make, feel, like, die, last, day, week, two, never, amount, time, get, happy, period, thing, happy, period, really, sure, actually, happy, get, guilty, happy, feel, weird, usually, manifest, ...]</td>\n",
              "      <td>even sure count depression year old girl struggle moderate anxiety early year old last year still struggle much since anxiety subside inconsistent bout depression tiniest thing trigger depression extremely intense come along depression get extremely angry lash isolate tend issue socially period really feel constant tiredness get tire easily wake start day feel need go back bed within hour feel destructive towards others think heavily suicide breakdown scream cry hair pull kind make feel like die last day week two never amount time get happy period thing happy period really sure actually happy get guilty happy feel weird usually manifest heavy feel chest stomach usually suicidal period social skill come back able enjoy thing feel muffle make sense overall issue motivation also tend punish hit head make sound crazy idk prescribed low dose lexapro question really need maybe unfortunate weird emotion fact depression inconsistent make feel like fake something talk brother suffers severe depression thought sound like numbness rather happiness fact able function enjoy thing slightly make feel otherwise stress tremendously make anxious inconsistent depression</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>210993</th>\n",
              "      <td>depression</td>\n",
              "      <td>[suffer, depression, tried, many, potential, solutions, medication, geographical, solutions, changing, locations, traveling, new, jobs, failure, presently, luckily, purely, liquid, assets, cash, stocks, options, etfs, bitcoin, etc, home, anything, value, debt, years, work, experience, engineering, industry, canadian, open, ideas, spend, money, general, maximum, potential, hopefully, one, day, get, depressive, state, would, presently, plan, travel, planet, money, runs, leave, planet, entirely, serious, would, depression]</td>\n",
              "      <td>[suffer, depression, try, many, potential, solution, medication, geographical, solution, change, location, travel, new, job, failure, presently, luckily, purely, liquid, asset, cash, stock, option, etf, bitcoin, etc, home, anything, value, debt, year, work, experience, engineering, industry, canadian, open, idea, spend, money, general, maximum, potential, hopefully, one, day, get, depressive, state, would, presently, plan, travel, planet, money, run, leave, planet, entirely, serious, would, depression]</td>\n",
              "      <td>suffer depression try many potential solution medication geographical solution change location travel new job failure presently luckily purely liquid asset cash stock option etf bitcoin etc home anything value debt year work experience engineering industry canadian open idea spend money general maximum potential hopefully one day get depressive state would presently plan travel planet money run leave planet entirely serious would depression</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "             label  ...                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                tokens_back_to_text\n",
              "129582  depression  ...                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     hi guy maybe help think depressed life never thought depression cause always thought feeling normal pretty sure use cry every single day every little thing would make cry fourth grade still know like extremely sensitive never care school make never homework pretty much life care could drive social anxiety think start parent tell useless would never amount anything never thought suicide always negative thought everything never remember ever happy im kind still like except way less severe anyone experienced thing thank advance think depressed since child\n",
              "199105  depression  ...                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    miserable life envy people life\n",
              "299657  depression  ...                                                                                                                                                                                                                                                                                                                                                                                                                                                                                          people may see perfectly normal person see deform monster see matter wear body fear smell im hideous expect people run ive treat like freak nature vulture picked away flaw im disgust entirety body im sure society become critical people appearance im treat like never make cut demon treat apart anymore year self confident ive make feel worthless unlovable unattractive year find difficult function society im constantly shot treat like im shadow ive finally fall rock bottom blazing molten core get home everyday scream mirror sit silence pas exhaustion repeat day go long im longer alive im zombie shuffle way distort reality birthday week dont want see time finally slip void finally time let go\n",
              "462176     bipolar  ...                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     im single year certain point life dont relationship nothing else matter almost reach point realize want find one continue work work find well job prospect constantly improve seem moot hard go unloved dramatic im feel many meme subreddit doesnt really feel genuine lonely\n",
              "196476  depression  ...                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               online relationship love life problem able tell upset want go numb upset talk asks want never say go upset met world warcraft play ton game together lot fun miss family though ocd make even harder tell want family feel like upset want alone would ruin know could right hard short ca tell people really feel due scar upset slightest way new could use advice\n",
              "125810  depression  ...  even sure count depression year old girl struggle moderate anxiety early year old last year still struggle much since anxiety subside inconsistent bout depression tiniest thing trigger depression extremely intense come along depression get extremely angry lash isolate tend issue socially period really feel constant tiredness get tire easily wake start day feel need go back bed within hour feel destructive towards others think heavily suicide breakdown scream cry hair pull kind make feel like die last day week two never amount time get happy period thing happy period really sure actually happy get guilty happy feel weird usually manifest heavy feel chest stomach usually suicidal period social skill come back able enjoy thing feel muffle make sense overall issue motivation also tend punish hit head make sound crazy idk prescribed low dose lexapro question really need maybe unfortunate weird emotion fact depression inconsistent make feel like fake something talk brother suffers severe depression thought sound like numbness rather happiness fact able function enjoy thing slightly make feel otherwise stress tremendously make anxious inconsistent depression\n",
              "210993  depression  ...                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                       suffer depression try many potential solution medication geographical solution change location travel new job failure presently luckily purely liquid asset cash stock option etf bitcoin etc home anything value debt year work experience engineering industry canadian open idea spend money general maximum potential hopefully one day get depressive state would presently plan travel planet money run leave planet entirely serious would depression\n",
              "\n",
              "[7 rows x 4 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 5
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nDdzvsHcMTMl",
        "outputId": "6bc4e9f3-afd6-47d6-80c2-914a36218602"
      },
      "source": [
        "df.isna().sum()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "label                  0\n",
              "text                   0\n",
              "lemmatized             0\n",
              "tokens_back_to_text    0\n",
              "dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0MsfZsxGZZsr",
        "outputId": "e8011709-53ec-4621-9d32-2612c08118e2"
      },
      "source": [
        "reddits = list(df['tokens_back_to_text'])\n",
        "type(reddits)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "list"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "orF09VH4Zg9Y",
        "outputId": "9e286102-0b31-44c4-d051-6fe6645da122"
      },
      "source": [
        "len(reddits)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "301746"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vpwv6by5n27I",
        "outputId": "0d3887e3-3fb8-4175-b4a6-8970031807f0"
      },
      "source": [
        "df['label'].value_counts()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "depression    178453\n",
              "Anxiety        84001\n",
              "bipolar        39292\n",
              "Name: label, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yO9yxGL5FIop"
      },
      "source": [
        "df['label'] = np.where(df['label']=='depression',1,0)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 890
        },
        "id": "k_xvvIBwn57v",
        "outputId": "1a2447eb-714a-4036-e971-a33c93ad15b5"
      },
      "source": [
        "df.head(3)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>label</th>\n",
              "      <th>text</th>\n",
              "      <th>lemmatized</th>\n",
              "      <th>tokens_back_to_text</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>0</td>\n",
              "      <td>[struggled, social, anxiety, childhood, main, advice, friends, books, professionals, expose, environments, better, example, get, anxious, every, single, time, go, shop, even, though, shopping, weekly, since, get, nervous, sweaty, heart, pounding, every, time, go, gym, yet, working, gym, years, exposure, worked, anyone, else, exposure, work]</td>\n",
              "      <td>[struggle, social, anxiety, childhood, main, advice, friend, book, professional, expose, environment, well, example, get, anxious, every, single, time, go, shop, even, though, shopping, weekly, since, get, nervous, sweaty, heart, pound, every, time, go, gym, yet, work, gym, year, exposure, work, anyone, else, exposure, work]</td>\n",
              "      <td>struggle social anxiety childhood main advice friend book professional expose environment well example get anxious every single time go shop even though shopping weekly since get nervous sweaty heart pound every time go gym yet work gym year exposure work anyone else exposure work</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>0</td>\n",
              "      <td>[back, march, know, ago, woke, day, go, school, feeling, kind, felt, depressed, really, feel, like, getting, even, seeing, anyone, history, p, really, mainly, happens, winter, time, get, highly, anxious, also, going, anywhere, periods, time, get, suspicious, paranoid, everyone, anything, usually, end, isolating, room, eat, way, less, also, get, highly, irritable, emotional, like, one, time, mom, sent, store, milk, started, crying, store, thought, going, killed, watched, believe, periods, government, something, higher, power, gerbil, studies, something, like, thought, news, real, used, put, fear, people, also, sort, figures, imagination, trying, manipulate, usually, happens, anywhere, day, weekish, also, insomniac, means, anything, p, anyways, yes, ...]</td>\n",
              "      <td>[back, march, know, ago, woke, day, go, school, feel, kind, felt, depressed, really, feel, like, get, even, see, anyone, history, p, really, mainly, happens, winter, time, get, highly, anxious, also, go, anywhere, period, time, get, suspicious, paranoid, everyone, anything, usually, end, isolate, room, eat, way, less, also, get, highly, irritable, emotional, like, one, time, mom, sent, store, milk, start, cry, store, thought, go, kill, watch, believe, period, government, something, high, power, gerbil, study, something, like, thought, news, real, use, put, fear, people, also, sort, figure, imagination, try, manipulate, usually, happens, anywhere, day, weekish, also, insomniac, mean, anything, p, anyways, yes, ...]</td>\n",
              "      <td>back march know ago woke day go school feel kind felt depressed really feel like get even see anyone history p really mainly happens winter time get highly anxious also go anywhere period time get suspicious paranoid everyone anything usually end isolate room eat way less also get highly irritable emotional like one time mom sent store milk start cry store thought go kill watch believe period government something high power gerbil study something like thought news real use put fear people also sort figure imagination try manipulate usually happens anywhere day weekish also insomniac mean anything p anyways yes thats background sit class feel highly anxious everything everyone guess get break point math class second block talk friend remember lot bit blurry happen explain real knew everything figure plan subject everything told thing life matter go die might die right extremely loud talk confuse unaware behaving teacher sat back class like nothing happen kid kind confuse kind explain understand alive point die one remember u really odd recap event suicidal really character end skip rest afternoon day friend request go home sleep bit talk friend happen tell scar go kill confuse despite talk loud much emotion besides try talk parent talk mom feeling nothing real say kid feel like time discus amongst friend ca relate couple month later mom take family doctor ask question kept say everyone felt really awkward find insomniac try go doctor grandma month two ask say felt paranoid sometimes call crazy ask eczema instead say talk know awkward go doctor discus really start affect personal school life thing happen never point school affect go store know sorry wall text start happen maybe year ago year first start year old advice appreciate panic attack derealization ca go doctor advice appreciate</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>0</td>\n",
              "      <td>[withdrawing, medicines, lately, effexor, zyprexa, cymbalta, time, mind, close, together, withdraw, thursday, major, panic, attack, bad, needed, take, extra, kolonopin, skip, work, friday, lasted, shaky, felt, sick, rest, night, saturday, quite, two, full, days, panic, attack, started, still, feel, like, describe, chihuahua, like, much, caffeine, except, small, amount, chocolate, yesterday, feels, like, edge, panic, attack, since, initial, one, sure, long, panic, attack, last]</td>\n",
              "      <td>[withdraw, medicine, lately, effexor, zyprexa, cymbalta, time, mind, close, together, withdraw, thursday, major, panic, attack, bad, need, take, extra, kolonopin, skip, work, friday, last, shaky, felt, sick, rest, night, saturday, quite, two, full, day, panic, attack, start, still, feel, like, describe, chihuahua, like, much, caffeine, except, small, amount, chocolate, yesterday, feel, like, edge, panic, attack, since, initial, one, sure, long, panic, attack, last]</td>\n",
              "      <td>withdraw medicine lately effexor zyprexa cymbalta time mind close together withdraw thursday major panic attack bad need take extra kolonopin skip work friday last shaky felt sick rest night saturday quite two full day panic attack start still feel like describe chihuahua like much caffeine except small amount chocolate yesterday feel like edge panic attack since initial one sure long panic attack last</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   label  ...                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                      tokens_back_to_text\n",
              "0      0  ...                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                struggle social anxiety childhood main advice friend book professional expose environment well example get anxious every single time go shop even though shopping weekly since get nervous sweaty heart pound every time go gym yet work gym year exposure work anyone else exposure work\n",
              "1      0  ...  back march know ago woke day go school feel kind felt depressed really feel like get even see anyone history p really mainly happens winter time get highly anxious also go anywhere period time get suspicious paranoid everyone anything usually end isolate room eat way less also get highly irritable emotional like one time mom sent store milk start cry store thought go kill watch believe period government something high power gerbil study something like thought news real use put fear people also sort figure imagination try manipulate usually happens anywhere day weekish also insomniac mean anything p anyways yes thats background sit class feel highly anxious everything everyone guess get break point math class second block talk friend remember lot bit blurry happen explain real knew everything figure plan subject everything told thing life matter go die might die right extremely loud talk confuse unaware behaving teacher sat back class like nothing happen kid kind confuse kind explain understand alive point die one remember u really odd recap event suicidal really character end skip rest afternoon day friend request go home sleep bit talk friend happen tell scar go kill confuse despite talk loud much emotion besides try talk parent talk mom feeling nothing real say kid feel like time discus amongst friend ca relate couple month later mom take family doctor ask question kept say everyone felt really awkward find insomniac try go doctor grandma month two ask say felt paranoid sometimes call crazy ask eczema instead say talk know awkward go doctor discus really start affect personal school life thing happen never point school affect go store know sorry wall text start happen maybe year ago year first start year old advice appreciate panic attack derealization ca go doctor advice appreciate\n",
              "2      0  ...                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                    withdraw medicine lately effexor zyprexa cymbalta time mind close together withdraw thursday major panic attack bad need take extra kolonopin skip work friday last shaky felt sick rest night saturday quite two full day panic attack start still feel like describe chihuahua like much caffeine except small amount chocolate yesterday feel like edge panic attack since initial one sure long panic attack last\n",
              "\n",
              "[3 rows x 4 columns]"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dNAB_rTNY58t",
        "outputId": "c1eba15d-daab-4a1e-d2df-0563711560c5"
      },
      "source": [
        "df['label'].value_counts()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1    178453\n",
              "0    123293\n",
              "Name: label, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fPwPfWWnprr0"
      },
      "source": [
        "## **Feature extraction**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YZpKi4zjptza"
      },
      "source": [
        "# prepare tokenizer\n",
        "tokenizer = Tokenizer()\n",
        "tokenizer.fit_on_texts(reddits)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1AIhZA8pp0wY",
        "outputId": "160947a7-04a2-4cc6-e15f-a05da292ebdd"
      },
      "source": [
        "word_index = tokenizer.word_index\n",
        "print('WE HAVE %s UNIQUE TOKENS.' % len(word_index))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "WE HAVE 95809 UNIQUE TOKENS.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2cXJjNzyp426"
      },
      "source": [
        "sequences = tokenizer.texts_to_sequences(reddits)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Juj92C51p7zA",
        "outputId": "12d834c8-d7d1-4d53-b45b-91c80e517090"
      },
      "source": [
        "# pad redddits to a max length of 150 words\n",
        "maxlen = 150\n",
        "X = pad_sequences(sequences, maxlen = maxlen)\n",
        "X.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(301746, 150)"
            ]
          },
          "metadata": {},
          "execution_count": 16
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q2Oc3_xbDwWk"
      },
      "source": [
        "y = df['label']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tZo2Np9bo696"
      },
      "source": [
        "##Split the dataset into **training** and **test** sets"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "17gtbAQvom9z"
      },
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=43)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "q0iVVvCVo-H4",
        "outputId": "ee76dde1-ec25-4aed-bdcf-fcbe91672ddf"
      },
      "source": [
        "X_train.shape, X_test.shape, y_train.shape, y_test.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((241396, 150), (60350, 150), (241396,), (60350,))"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t5DtzExcrPuV"
      },
      "source": [
        "y_train = np.asarray(y_train).astype('float32').reshape((-1,1))\n",
        "y_test = np.asarray(y_test).astype('float32').reshape((-1,1))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "24wLyGZApa_B"
      },
      "source": [
        "# Defining the **Convolutional Neural Network Model**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lO9o6yzwpUIi"
      },
      "source": [
        "# Prepare hyperparameters of CNN\n",
        "vocab_size = len(word_index) + 1\n",
        "# output_dim = 32"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zrk1T7vWpium"
      },
      "source": [
        "# Building the CNN Model\n",
        "model = Sequential()\n",
        "\n",
        "model.add(Embedding(vocab_size, 128, input_length=maxlen))\n",
        "model.add(Conv1D(128, 3, padding='same', activation='relu'))\n",
        "model.add(Conv1D(64, 3, padding='same', activation='relu'))\n",
        "model.add(MaxPooling1D())\n",
        "model.add(Flatten())\n",
        "\n",
        "model.add(Dense(250, activation='relu'))\n",
        "\n",
        "model.add(Dense(1, activation='sigmoid'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sDQVHW1pEPm4"
      },
      "source": [
        "### Visualizing the CNN Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 758
        },
        "id": "bbF56xATLULt",
        "outputId": "46ad398f-6c5a-4a00-86fe-4739b4db89ce"
      },
      "source": [
        "plot_model(model)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2AXhVEpKqLsg",
        "outputId": "7235f515-3b7b-4828-e44e-2bd1bb744f5c"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "embedding (Embedding)        (None, 150, 128)          12263680  \n",
            "_________________________________________________________________\n",
            "conv1d (Conv1D)              (None, 150, 128)          49280     \n",
            "_________________________________________________________________\n",
            "conv1d_1 (Conv1D)            (None, 150, 64)           24640     \n",
            "_________________________________________________________________\n",
            "max_pooling1d (MaxPooling1D) (None, 75, 64)            0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 4800)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 250)               1200250   \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 1)                 251       \n",
            "=================================================================\n",
            "Total params: 13,538,101\n",
            "Trainable params: 13,538,101\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MtvNxeoiEiTH"
      },
      "source": [
        "### Using the **Adam optimizer** for the optimization and **Binary cross entropy** to calculate the Loss Function"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z1pXLbl_LzV-"
      },
      "source": [
        "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "o7ZoCd_Qq0Vp",
        "outputId": "d879322c-7394-4a9f-8d5e-ceb2751eb2c9"
      },
      "source": [
        "start = time.time()\n",
        "\n",
        "# fit the model to the training set\n",
        "cnn_hist = model.fit(X_train, y_train, epochs = 5, validation_split = 0.2)\n",
        "\n",
        "# evaluation of the model on the test set\n",
        "scores = model.evaluate(X_test, y_test, verbose = 0)\n",
        "print(\"Accuracy: %.2f%%\" % (scores[1]*100))\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "\n",
        "end = time.time()\n",
        "total = (end-start) // 60\n",
        "print('Training duration : {} minutes'.format(total))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/5\n",
            "6035/6035 [==============================] - 1187s 197ms/step - loss: 0.3080 - accuracy: 0.8744 - val_loss: 0.2852 - val_accuracy: 0.8846\n",
            "Epoch 2/5\n",
            "6035/6035 [==============================] - 1184s 196ms/step - loss: 0.2600 - accuracy: 0.8976 - val_loss: 0.2843 - val_accuracy: 0.8866\n",
            "Epoch 3/5\n",
            "6035/6035 [==============================] - 1188s 197ms/step - loss: 0.2210 - accuracy: 0.9143 - val_loss: 0.3104 - val_accuracy: 0.8783\n",
            "Epoch 4/5\n",
            "6035/6035 [==============================] - 1187s 197ms/step - loss: 0.1758 - accuracy: 0.9333 - val_loss: 0.3427 - val_accuracy: 0.8748\n",
            "Epoch 5/5\n",
            "6035/6035 [==============================] - 1193s 198ms/step - loss: 0.1310 - accuracy: 0.9507 - val_loss: 0.4454 - val_accuracy: 0.8640\n",
            "Accuracy: 86.51%\n",
            "Training duration : 99.0 minutes\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b9UZy78yEK_U"
      },
      "source": [
        "history_dict = cnn_hist.history"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gutqv5GsKG0W"
      },
      "source": [
        "## **Plotting the loss and the accuracy of both training and validation sets using the model history**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u0R9WAp63esX",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 513
        },
        "outputId": "2883ae18-8577-4dcb-d225-c5a581eabf0f"
      },
      "source": [
        "# loss\n",
        "loss_values = history_dict['loss']\n",
        "val_loss_values = history_dict['val_loss']\n",
        "\n",
        "epochs = range(1, len(loss_values) + 1) \n",
        "\n",
        "# plot\n",
        "plt.figure(figsize=(12, 8))\n",
        "plt.plot(epochs, loss_values, 'blue',label='Training loss')\n",
        "plt.plot(epochs, val_loss_values, 'orange', label='Validation loss')\n",
        "plt.title('Training and validation loss')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c26VWaqPFfQ7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 513
        },
        "outputId": "a8c043bf-4912-48b9-989b-5af5f7eaba62"
      },
      "source": [
        "# accuracy\n",
        "acc = history_dict['accuracy']\n",
        "val_acc = history_dict['val_accuracy']\n",
        "\n",
        "# range of X (no. of epochs)\n",
        "epochs = range(1, len(acc) + 1)\n",
        "\n",
        "plt.figure(figsize=(12, 8))\n",
        "plt.plot(epochs, acc, 'blue',label='Training accuracy')\n",
        "# orange is for \"orange\"\n",
        "plt.plot(epochs, val_acc, 'orange', label='Validation accuracy')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gpQDITnpJy7X"
      },
      "source": [
        "## **Confusion matrix**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fdm2im9sFFd4"
      },
      "source": [
        "preds = np.round(model.predict(X_test),0)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eoDqL9Xj9FMj"
      },
      "source": [
        "cm = confusion_matrix(y_test, preds)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yZxhKaOV_aAa"
      },
      "source": [
        "df_cm = pd.DataFrame(cm, index = [\"depression\", \"non-depression\"], columns = [\"depression\", \"non-depression\"])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 302
        },
        "id": "uAfhW45D8sQI",
        "outputId": "450a02c1-ee27-4fe2-d0c0-683818980909"
      },
      "source": [
        "plt.figure(figsize=(6, 4))\n",
        "sb.heatmap(df_cm, annot=True, cmap=\"Reds\", fmt='d')\n",
        "plt.xlabel('True Class')\n",
        "plt.ylabel('Predicted Class')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(30.499999999999993, 0.5, 'Predicted Class')"
            ]
          },
          "metadata": {},
          "execution_count": 81
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iggsDHPH8tdS"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}