petite-difference-challenge2/dev-0/model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "5b55a105",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "id": "9364cf2c",
   "metadata": {},
   "outputs": [],
   "source": [
    "tsv_data = pd.read_csv('in.tsv', sep='\\t',header=None, quoting=csv.QUOTE_NONE)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "id": "9d3f7db9",
   "metadata": {},
   "outputs": [],
   "source": [
    "expected = pd.read_csv('expected.tsv', sep='\\t',header=None)[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "5062478d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "137314\n",
      "137314\n"
     ]
    }
   ],
   "source": [
    "print(len(expected))\n",
    "print(len(tsv_data))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "id": "5eca7aab",
   "metadata": {},
   "outputs": [],
   "source": [
    "male={'silnik', 'windows', 'gb', 'mb', 'mecz', 'pc', 'opony', 'apple', 'iphone', 'zwiastuny', 'hd', 'ubuntu', 'system', 'serwer'}\n",
    "female={'ciąża', 'miesiączki', 'ciasto', 'ciąże', 'zadowolona', 'antykoncepcyjne', 'ginekologia', 'tabletki', 'porodzie', 'mąż', 'krwawienie', 'ciasta'}\n",
    "male = {x[:6].lower() for x in male}\n",
    "female = {x[:6].lower() for x in female}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "id": "0bdd1845",
   "metadata": {},
   "outputs": [],
   "source": [
    "trimmed_docs=[]\n",
    "for document in tsv_data:\n",
    "    new_doc=[]\n",
    "    for word in str(document).lower().split():\n",
    "        new_doc.append(word[:6])\n",
    "    trimmed_docs.append(new_doc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "id": "b36bbd92",
   "metadata": {},
   "outputs": [],
   "source": [
    "male_or_female=[]\n",
    "for doc in trimmed_docs:\n",
    "    male_or_female.append((len(male&set(doc)), len(female&set(doc))))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "id": "ccbad95c",
   "metadata": {},
   "outputs": [],
   "source": [
    "answers=[]\n",
    "for i in male_or_female:\n",
    "    if i[0]>i[1]:\n",
    "        answers.append(1)\n",
    "    else:\n",
    "        answers.append(0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "id": "02ee0acf",
   "metadata": {},
   "outputs": [],
   "source": [
    "result=[]\n",
    "for i in range(len(answers)):\n",
    "    if answers[i]==expected[i]:\n",
    "        result.append(1)\n",
    "    else:\n",
    "        result.append(0)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "id": "db803a58",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predykcja modelu wynosi 51.007909%\n"
     ]
    }
   ],
   "source": [
    "print(f'Predykcja modelu wynosi {sum(result)/len(result)*100:.6f}%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "id": "e1a15db7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['cierpi',\n",
       " 'na',\n",
       " 'strasz',\n",
       " 'lagi',\n",
       " '–',\n",
       " 'kilkan',\n",
       " 'sekund',\n",
       " 'lub',\n",
       " 'dłużej',\n",
       " 'czarne',\n",
       " 'ekranu',\n",
       " 'przy',\n",
       " 'próbie',\n",
       " 'przełą',\n",
       " 'się',\n",
       " '/',\n",
       " 'urucho',\n",
       " 'prawie',\n",
       " 'każdej',\n",
       " 'aplika',\n",
       " 'dodatk',\n",
       " 'telefo',\n",
       " 'mi',\n",
       " 'się',\n",
       " 'wyłącz',\n",
       " 'czasem',\n",
       " 'bez',\n",
       " 'powodu',\n",
       " '–',\n",
       " 'sam',\n",
       " 'z',\n",
       " 'siebie',\n",
       " 'albo',\n",
       " 'reseto',\n",
       " 'ostatn',\n",
       " 'nawet',\n",
       " 'przegl',\n",
       " 'zaczęł',\n",
       " 'się',\n",
       " 'często',\n",
       " 'zawies',\n",
       " 'i',\n",
       " 'androi',\n",
       " 'propon',\n",
       " 'wymusz',\n",
       " 'zamkni',\n",
       " 'do',\n",
       " 'tego',\n",
       " 'te',\n",
       " 'proble',\n",
       " 'z',\n",
       " 'połącz',\n",
       " 'do',\n",
       " 'komput',\n",
       " 'przez',\n",
       " 'usb.']"
      ]
     },
     "execution_count": 164,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trimmed_docs[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7403c1bb",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}