Assignments/a4/written/BLEU_Verify.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# BLEU Varify"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from nltk.translate.bleu_score import sentence_bleu\n",
    "import numpy as np\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\") # ignore nltk warnings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "references = ['love can always find a way'.split(),\n",
    "              'love makes anything possible'.split()]\n",
    "candidate1 = 'the love can always do'.split()\n",
    "candidate2 = 'love can make anything possible'.split()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": "0.6\n0.8\n"
    }
   ],
   "source": [
    "# Candidate1 unigram\n",
    "score = sentence_bleu(references, candidate1, weights=(1, 0, 0, 0))\n",
    "print(score)\n",
    "# Candidate2 unigram\n",
    "score = sentence_bleu(references, candidate2, weights=(1, 0, 0, 0))\n",
    "print(score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": "0.5\n0.5\n"
    }
   ],
   "source": [
    "# Candidate1 bigram\n",
    "score = sentence_bleu(references, candidate1, weights=(0, 1, 0, 0))\n",
    "print(score)\n",
    "# Candidate2 bigram\n",
    "score = sentence_bleu(references, candidate2, weights=(0, 1, 0, 0))\n",
    "print(score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": "0.5477225575051662\n0.6324555320336759\n"
    }
   ],
   "source": [
    "# Candidate1 unigram + bigram\n",
    "score = sentence_bleu(references, candidate1, weights=(0.5, 0.5, 0, 0))\n",
    "print(score)\n",
    "# Candidate2 unigram + bigram\n",
    "score = sentence_bleu(references, candidate2, weights=(0.5, 0.5, 0, 0))\n",
    "print(score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "reference1 = ['love can always find a way'.split()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": "0.49123845184678916\n0.3274923012311928\n"
    }
   ],
   "source": [
    "# Candidate1 unigram\n",
    "score = sentence_bleu(reference1, candidate1, weights=(1, 0, 0, 0))\n",
    "print(score)\n",
    "# Candidate2 unigram\n",
    "score = sentence_bleu(reference1, candidate2, weights=(1, 0, 0, 0))\n",
    "print(score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": "0.40936537653899097\n0.20468268826949548\n"
    }
   ],
   "source": [
    "# Candidate1 bigram\n",
    "score = sentence_bleu(reference1, candidate1, weights=(0, 1, 0, 0))\n",
    "print(score)\n",
    "# Candidate2 bigram\n",
    "score = sentence_bleu(reference1, candidate2, weights=(0, 1, 0, 0))\n",
    "print(score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": "0.448437301984003\n0.25890539701513365\n"
    }
   ],
   "source": [
    "# Candidate1 unigram + bigram\n",
    "score = sentence_bleu(reference1, candidate1, weights=(0.5, 0.5, 0, 0))\n",
    "print(score)\n",
    "# Candidate2 unigram + bigram\n",
    "score = sentence_bleu(reference1, candidate2, weights=(0.5, 0.5, 0, 0))\n",
    "print(score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": "0.5477225575051662"
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.exp((0.5 * np.log(0.6) + 0.5 * np.log(0.5)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": "0.6324555320336759"
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.exp((0.5 * np.log(0.8) + 0.5 * np.log(0.5)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": "0.8187307530779819"
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.exp(1 - 6/5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": "0.448437301984003"
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.exp(1 - 6/5) * np.exp((0.5 * np.log(0.6) + 0.5 * np.log(0.5)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": "0.25890539701513365"
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.exp(1 - 6/5) * np.exp((0.5 * np.log(0.4) + 0.5 * np.log(0.25)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.5"
  },
  "latex_envs": {
   "LaTeX_envs_menu_present": true,
   "autoclose": false,
   "autocomplete": true,
   "bibliofile": "biblio.bib",
   "cite_by": "apalike",
   "current_citInitial": 1,
   "eqLabelWithNumbers": true,
   "eqNumInitial": 1,
   "hotkeys": {
    "equation": "Ctrl-E",
    "itemize": "Ctrl-I"
   },
   "labels_anchors": false,
   "latex_user_defs": false,
   "report_style_numbering": false,
   "user_envs_cfg": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}