Evaluating_Approaches.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Evaluating my approach to target identification"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Notebook Aim:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> The aim of this notebook is to compare different approaches to identifying target players from a phrase of a football match report "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Outcome:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> Produce the best possible method for identifying the target player without increasing the number of false positives (identifying players that are not targets)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Methods:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> - The model text has been scraped from articles on the first page of 'The Mirror' match report site\n",
    "- It has been split into sentences and then split again by commas to produce short phrases\n",
    "- I will evaluate each method by comparing it to a manually annotated dataset of the same phrases\n",
    "- The accuracy of each method will be determined by the number of rows that identified the same target(s) as identified in the annotated dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "--------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Import the Annotated Data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Create a Dataframe from Model Phrases"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 326,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "pd.set_option('display.max_colwidth', -1)\n",
    "pd.set_option('display.max_rows', None)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 327,
   "metadata": {},
   "outputs": [],
   "source": [
    "annotated_df = pd.read_csv('annotated_data.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 328,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 329,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Replace NaN with empty string\n",
    "annotated_df = annotated_df.replace(np.nan, '', regex=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Extract Sentences Column for Target Identification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 330,
   "metadata": {},
   "outputs": [],
   "source": [
    "phrases_df = pd.DataFrame(annotated_df['Sentence'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "-----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Target Identification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 331,
   "metadata": {},
   "outputs": [],
   "source": [
    "import spacy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 332,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load spaCy model\n",
    "nlp = spacy.load(\"en_core_web_sm\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 333,
   "metadata": {},
   "outputs": [],
   "source": [
    "import targetevaluation_playernames"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 334,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Flatten list of player names\n",
    "players = [y for x in list(targetevaluation_playernames.player_names.values()) for y in x]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Approach 1 - Identify NSubj\n",
    "- Function will extract players that are the main subject of a sentence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 335,
   "metadata": {},
   "outputs": [],
   "source": [
    "def identify_nsubj(string):\n",
    "    \n",
    "    # Create spacy object from string\n",
    "    doc = nlp(string)\n",
    "    \n",
    "    # Create empty nsubj list\n",
    "    nsubj = []\n",
    "    \n",
    "    # Iterate through doc object\n",
    "    for i, tok in enumerate(doc):\n",
    "        # Check for NSubj only if tok is a proper noun and it's length is greater than 2\n",
    "        if str(tok.dep_) == \"nsubj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            # Apped subject if it matches with a player\n",
    "            nsubj.append([player for player in players if tok.text in player])\n",
    "    \n",
    "    # return a flattened list\n",
    "    return str([y for x in nsubj for y in x])[1:-1].replace(\"'\", \"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 336,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply function to phrases_df\n",
    "phrases_df['nsubj'] = phrases_df['Sentence'].apply(identify_nsubj)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Approach 2 - Identify NSubj and Subjpass\n",
    "- Function will extract players that are main subject of a sentence and passive subjects of a sentence "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 337,
   "metadata": {},
   "outputs": [],
   "source": [
    "def identify_nsubj_passsubj(string):\n",
    "    \n",
    "    # Create spacy object from string\n",
    "    doc = nlp(string)\n",
    "    \n",
    "    # Create empty nsubj list\n",
    "    nsubj_passsubj = []\n",
    "    \n",
    "    # Iterate through doc object\n",
    "    for i, tok in enumerate(doc):\n",
    "        # ADDING NSUBJ\n",
    "        # Check for NSubj only if tok is a proper noun and it's length is greater than 2\n",
    "        if str(tok.dep_) == \"nsubj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            # Apped subject if it matches with a player\n",
    "            nsubj_passsubj.append([player for player in players if tok.text in player])\n",
    "        # ADDING SUBJPASS\n",
    "        elif str(tok.dep_) == \"subjpass\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_passsubj.append([player for player in players if tok.text in player])\n",
    "            \n",
    "    # return a flattened list        \n",
    "    return str([y for x in nsubj_passsubj for y in x])[1:-1].replace(\"'\", \"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 338,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply function to phrases_df\n",
    "phrases_df['nsubj_passsubj'] = phrases_df['Sentence'].apply(identify_nsubj_passsubj)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Approach 3 - Identify Nsubj and Pobj\n",
    "- Function will extract players that are main subject of a sentence and objects of a preposition "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 339,
   "metadata": {},
   "outputs": [],
   "source": [
    "def identify_nsubj_pobj(string):\n",
    "    \n",
    "    # Create spacy object from string\n",
    "    doc = nlp(string)\n",
    "    \n",
    "    # Create empty nsubj list\n",
    "    nsubj_pobj = []\n",
    "    \n",
    "    # Iterate through doc object\n",
    "    for i, tok in enumerate(doc):\n",
    "        # ADDING NSUBJ\n",
    "        # Check for NSubj only if tok is a proper noun and it's length is greater than 2\n",
    "        if str(tok.dep_) == \"nsubj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            # Apped subject if it matches with a player\n",
    "            nsubj_pobj.append([player for player in players if tok.text in player])\n",
    "        # ADDING POBJ\n",
    "        elif str(tok.dep_) == \"pobj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_pobj.append([player for player in players if tok.text in player])\n",
    "            \n",
    "    # return a flattened list        \n",
    "    return str([y for x in nsubj_pobj for y in x])[1:-1].replace(\"'\", \"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 340,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply function to phrases_df\n",
    "phrases_df['nsubj_pobj'] = phrases_df['Sentence'].apply(identify_nsubj_pobj)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Approach 4 - Identify Nsubj, Pobj and Dobj\n",
    "- Function will extract players that are main subject of a sentence, objects of a preposition and direct object of a sentence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 341,
   "metadata": {},
   "outputs": [],
   "source": [
    "def identify_nsubj_pobj_dobj(string):\n",
    "    \n",
    "    # Create spacy object from string\n",
    "    doc = nlp(string)\n",
    "    \n",
    "    # Create empty nsubj list\n",
    "    nsubj_pobj_dobj = []\n",
    "    \n",
    "    # Iterate through doc object\n",
    "    for i, tok in enumerate(doc):\n",
    "        # ADDING NSUBJ\n",
    "        # Check for NSubj only if tok is a proper noun and it's length is greater than 2\n",
    "        if str(tok.dep_) == \"nsubj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            # Apped subject if it matches with a player\n",
    "            nsubj_pobj_dobj.append([player for player in players if tok.text in player])\n",
    "        # ADDING POBJ\n",
    "        elif str(tok.dep_) == \"pobj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_pobj_dobj.append([player for player in players if tok.text in player])\n",
    "        # ADDING DOBJ\n",
    "        elif str(tok.dep_) == \"dobj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_pobj_dobj.append([player for player in players if tok.text in player])\n",
    "            \n",
    "    # return a flattened list        \n",
    "    return str([y for x in nsubj_pobj_dobj for y in x])[1:-1].replace(\"'\", \"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 342,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply function to phrases_df\n",
    "phrases_df['nsubj_pobj_dobj'] = phrases_df['Sentence'].apply(identify_nsubj_pobj_dobj)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Approach 5 - Identify Nsubj, Dobj\n",
    "- Function will extract players that are main subject of a sentence and direct objects of a sentence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 343,
   "metadata": {},
   "outputs": [],
   "source": [
    "def identify_nsubj_dobj(string):\n",
    "    \n",
    "    # Create spacy object from string\n",
    "    doc = nlp(string)\n",
    "    \n",
    "    # Create empty nsubj list\n",
    "    nsubj_dobj = []\n",
    "    \n",
    "    # Iterate through doc object\n",
    "    for i, tok in enumerate(doc):\n",
    "        # ADDING NSUBJ\n",
    "        # Check for NSubj only if tok is a proper noun and it's length is greater than 2\n",
    "        if str(tok.dep_) == \"nsubj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            # Apped subject if it matches with a player\n",
    "            nsubj_dobj.append([player for player in players if tok.text in player])\n",
    "        # ADDING DOBJ\n",
    "        elif str(tok.dep_) == \"dobj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_dobj.append([player for player in players if tok.text in player])\n",
    "            \n",
    "    # return a flattened list        \n",
    "    return str([y for x in nsubj_dobj for y in x])[1:-1].replace(\"'\", \"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 344,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply function to phrases_df\n",
    "phrases_df['nsubj_dobj'] = phrases_df['Sentence'].apply(identify_nsubj_dobj)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Approach 6 - Identify Nsubj, Poss\n",
    "- Function will extract players that are main subject of a sentence and possessives"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 345,
   "metadata": {},
   "outputs": [],
   "source": [
    "def identify_nsubj_poss(string):\n",
    "    \n",
    "    # Create spacy object from string\n",
    "    doc = nlp(string)\n",
    "    \n",
    "    # Create empty nsubj list\n",
    "    nsubj_poss = []\n",
    "    \n",
    "    # Iterate through doc object\n",
    "    for i, tok in enumerate(doc):\n",
    "        # ADDING NSUBJ\n",
    "        # Check for NSubj only if tok is a proper noun and it's length is greater than 2\n",
    "        if str(tok.dep_) == \"nsubj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            # Apped subject if it matches with a player\n",
    "            nsubj_poss.append([player for player in players if tok.text in player])\n",
    "        # ADDING POSS\n",
    "        elif str(tok.dep_) == \"poss\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_poss.append([player for player in players if tok.text in player])\n",
    "            \n",
    "    # return a flattened list        \n",
    "    return str([y for x in nsubj_poss for y in x])[1:-1].replace(\"'\", \"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 346,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply function to phrases_df\n",
    "phrases_df['nsubj_poss'] = phrases_df['Sentence'].apply(identify_nsubj_poss)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Approach 7 - Identify Nsubj, Pobj, Dobj and Poss\n",
    "- Function will extract players that are main subject of a sentence, object of a preposition, direct object and possessives"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 347,
   "metadata": {},
   "outputs": [],
   "source": [
    "def identify_nsubj_pobj_dobj_poss(string):\n",
    "    \n",
    "    # Create spacy object from string\n",
    "    doc = nlp(string)\n",
    "    \n",
    "    # Create empty nsubj list\n",
    "    nsubj_pobj_dobj_poss = []\n",
    "    \n",
    "    # Iterate through doc object\n",
    "    for i, tok in enumerate(doc):\n",
    "        # ADDING NSUBJ\n",
    "        # Check for NSubj only if tok is a proper noun and it's length is greater than 2\n",
    "        if str(tok.dep_) == \"nsubj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            # Apped subject if it matches with a player\n",
    "            nsubj_pobj_dobj_poss.append([player for player in players if tok.text in player])\n",
    "        # ADDING POBJ\n",
    "        elif str(tok.dep_) == \"pobj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_pobj_dobj_poss.append([player for player in players if tok.text in player])\n",
    "        # ADDING DOBJ\n",
    "        elif str(tok.dep_) == \"dobj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_pobj_dobj_poss.append([player for player in players if tok.text in player])\n",
    "        # ADDING POSS\n",
    "        elif str(tok.dep_) == \"poss\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_pobj_dobj_poss.append([player for player in players if tok.text in player])\n",
    "            \n",
    "    # return a flattened list        \n",
    "    return str([y for x in nsubj_pobj_dobj_poss for y in x])[1:-1].replace(\"'\", \"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 348,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply function to phrases_df\n",
    "phrases_df['nsubj_pobj_dobj_poss'] = phrases_df['Sentence'].apply(identify_nsubj_pobj_dobj_poss)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Approach 8 - Identify Nsubj, Pobj and Poss\n",
    "- Function will extract players that are main subject of a sentence, object of a preposition and possessives"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 365,
   "metadata": {},
   "outputs": [],
   "source": [
    "def identify_nsubj_pobj_poss(string):\n",
    "    \n",
    "    # Create spacy object from string\n",
    "    doc = nlp(string)\n",
    "    \n",
    "    # Create empty nsubj_pobj_poss list\n",
    "    nsubj_pobj_poss = []\n",
    "    \n",
    "    # Iterate through doc object\n",
    "    for i, tok in enumerate(doc):\n",
    "        # ADDING NSUBJ\n",
    "        # Check for NSubj only if tok is a proper noun and it's length is greater than 2\n",
    "        if str(tok.dep_) == \"nsubj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            # Apped subject if it matches with a player\n",
    "            nsubj_pobj_poss.append([player for player in players if tok.text in player])\n",
    "        # ADDING POBJ\n",
    "        elif str(tok.dep_) == \"pobj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_pobj_poss.append([player for player in players if tok.text in player])\n",
    "        # ADDING POSS\n",
    "        elif str(tok.dep_) == \"poss\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_pobj_poss.append([player for player in players if tok.text in player])\n",
    "            \n",
    "    # return a flattened list        \n",
    "    return str([y for x in nsubj_pobj_poss for y in x])[1:-1].replace(\"'\", \"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 366,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply function to phrases_df\n",
    "phrases_df['nsubj_pobj_poss'] = phrases_df['Sentence'].apply(identify_nsubj_pobj_poss)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Approach 9 - Identify Nsubj, Dobj, Poss\n",
    "- Function will extract players that are main subject of a sentence, direct object and possessives"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 367,
   "metadata": {},
   "outputs": [],
   "source": [
    "def identify_nsubj_dobj_poss(string):\n",
    "    \n",
    "    # Create spacy object from string\n",
    "    doc = nlp(string)\n",
    "    \n",
    "    # Create empty nsubj_pobj_poss list\n",
    "    nsubj_dobj_poss = []\n",
    "    \n",
    "    # Iterate through doc object\n",
    "    for i, tok in enumerate(doc):\n",
    "        # ADDING NSUBJ\n",
    "        # Check for NSubj only if tok is a proper noun and it's length is greater than 2\n",
    "        if str(tok.dep_) == \"nsubj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            # Apped subject if it matches with a player\n",
    "            nsubj_dobj_poss.append([player for player in players if tok.text in player])\n",
    "        # ADDING POBJ\n",
    "        elif str(tok.dep_) == \"dobj\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_dobj_poss.append([player for player in players if tok.text in player])\n",
    "        # ADDING POSS\n",
    "        elif str(tok.dep_) == \"poss\" and str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            nsubj_dobj_poss.append([player for player in players if tok.text in player])\n",
    "            \n",
    "    # return a flattened list        \n",
    "    return str([y for x in nsubj_dobj_poss for y in x])[1:-1].replace(\"'\", \"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 368,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply function to phrases_df\n",
    "phrases_df['nsubj_dobj_poss'] = phrases_df['Sentence'].apply(identify_nsubj_dobj_poss)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Approach 10 - Identify any mention of a player\n",
    "- Function will extract any player that is mentioned in the sentence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 349,
   "metadata": {},
   "outputs": [],
   "source": [
    "def identify_mention(string):\n",
    "    \n",
    "    # Create spacy object from string\n",
    "    doc = nlp(string)\n",
    "    \n",
    "    # Create empty nsubj list\n",
    "    mention = []\n",
    "    \n",
    "    # Iterate through doc object\n",
    "    for i, tok in enumerate(doc):\n",
    "        # Check if tok is a proper noun and it's length is greater than 2\n",
    "        if str(tok.pos_) == \"PROPN\" and len(tok.text) > 2:\n",
    "            # Apped subject if it matches with a player\n",
    "            mention.append([player for player in players if tok.text in player])\n",
    "            \n",
    "    \n",
    "    \n",
    "    # return a flattened list\n",
    "    return str(list(dict.fromkeys([y.strip(\"''\") for x in mention for y in x])))[1:-1].replace(\"'\", \"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 350,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply function to phrases_df\n",
    "phrases_df['mention'] = phrases_df['Sentence'].apply(identify_mention)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Compare accuracy of approaches"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ##### Function to calculate accuracy score of approach will:\n",
    "- Compare common rows between annotated dataset and column in phrases df\n",
    "- Produce a confusion score for false positives, false negatives, true negatives, true positives"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Accuracy Score = `(correct matches / length of dataframe) * 100`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 476,
   "metadata": {},
   "outputs": [],
   "source": [
    "def accuracy(df_column):\n",
    "    \n",
    "    # Accuracy counter \n",
    "    accuracy_counter = 0\n",
    "    \n",
    "    # Confusion score dictionary\n",
    "    confusion_dictionary = {\n",
    "        \"Approach\": df_column,\n",
    "        \"True Positive\": 0,\n",
    "        \"True Negative\": 0,\n",
    "        \"False Positive\": 0,\n",
    "        \"False Negative\": 0\n",
    "    }\n",
    "    \n",
    "    # ====================\n",
    "    # ACCURACY SCORE \n",
    "    # ====================\n",
    "    \n",
    "    # Iterate through phrases_df & annotated_df\n",
    "    for index, row in phrases_df.iterrows():\n",
    "        # Increment Accuracy Score\n",
    "        if row[df_column] == annotated_df.iloc[index][\"Target\"]:\n",
    "            accuracy_counter += 1\n",
    "        else: \n",
    "            continue\n",
    "    \n",
    "    \n",
    "    for index, row in phrases_df.iterrows():\n",
    "        # ====================\n",
    "        # CONFUSION DICTIONARY \n",
    "        # ====================\n",
    "        \n",
    "        # If targets are equal\n",
    "        if row[df_column] == annotated_df.iloc[index][\"Target\"]:\n",
    "            # Increment True Positive\n",
    "            if annotated_df.iloc[index][\"Target\"] != \"\":\n",
    "                confusion_dictionary[\"True Positive\"] += 1\n",
    "            # Increment True Negative\n",
    "            elif annotated_df.iloc[index][\"Target\"] == \"\":\n",
    "                confusion_dictionary[\"True Negative\"] += 1\n",
    "            # Print Error\n",
    "            else: \n",
    "                print(\"Error 1: Conditions don't catch all cases\")\n",
    "        elif row[df_column] != annotated_df.iloc[index][\"Target\"]:\n",
    "            # Increment False Positive\n",
    "            if annotated_df.iloc[index][\"Target\"] == \"\":\n",
    "                confusion_dictionary[\"False Positive\"] += 1\n",
    "            # Increment False Negative\n",
    "            elif annotated_df.iloc[index][\"Target\"] != \"\" and row[df_column] == \"\":\n",
    "                confusion_dictionary[\"False Negative\"] += 1\n",
    "            # If Neither are blank but they don't match ---> Split the two columns into lists and work out which is longer\n",
    "            else:\n",
    "                # If the annotated entry is longer then its a false negative because targets weren't recognised\n",
    "                if len(annotated_df.iloc[index][\"Target\"].split(',')) > len(row[df_column].split(',')):\n",
    "                    confusion_dictionary[\"False Negative\"] += 1\n",
    "                # If the phrases_df entry is longer then its a false positive because targets were recognised that weren't there\n",
    "                elif len(annotated_df.iloc[index][\"Target\"].split(',')) < len(row[df_column].split(',')):\n",
    "                    confusion_dictionary[\"False Positive\"] += 1\n",
    "                # If the they match in length its a false positive because the entries weren't the same\n",
    "                elif len(annotated_df.iloc[index][\"Target\"].split(',')) == len(row[df_column].split(',')):\n",
    "                    confusion_dictionary[\"False Positive\"] += 1\n",
    "                else: \n",
    "                    print(\"Error 2: Conditions don't catch all cases\")\n",
    "                \n",
    "        else: \n",
    "            print(\"Error 3: Conditions don't catch all cases\")\n",
    "            \n",
    "        \n",
    "        \n",
    "            \n",
    "            \n",
    "    \n",
    "    # Calculate Accuracy Score\n",
    "    accuracy_score = (accuracy_counter / len(phrases_df)) * 100\n",
    "    \n",
    "    \n",
    "    return  round(accuracy_score, 3), confusion_dictionary"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Test Accuracy - Approach 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 477,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save Variables\n",
    "accuracy_1, confusion_1 = accuracy(\"nsubj\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 478,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Score: 84.97 {'Approach': 'nsubj', 'True Positive': 40, 'True Negative': 384, 'False Positive': 3, 'False Negative': 72}\n"
     ]
    }
   ],
   "source": [
    "# Print Variables\n",
    "print(\"Accuracy Score:\", accuracy_1, confusion_1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Test Accuracy - Approach 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 479,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save Variables\n",
    "accuracy_2, confusion_2 = accuracy(\"nsubj_passsubj\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 480,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Score: 84.97 {'Approach': 'nsubj_passsubj', 'True Positive': 40, 'True Negative': 384, 'False Positive': 3, 'False Negative': 72}\n"
     ]
    }
   ],
   "source": [
    "# Print Variables\n",
    "print(\"Accuracy Score:\", accuracy_2, confusion_2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Test Accuracy - Approach 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 481,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save Variables\n",
    "accuracy_3, confusion_3 = accuracy(\"nsubj_pobj\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 482,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Score: 88.978 {'Approach': 'nsubj_pobj', 'True Positive': 64, 'True Negative': 380, 'False Positive': 10, 'False Negative': 45}\n"
     ]
    }
   ],
   "source": [
    "# Print Variables\n",
    "print(\"Accuracy Score:\", accuracy_3, confusion_3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Test Accuracy - Approach 4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 483,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save Variables\n",
    "accuracy_4, confusion_4 = accuracy(\"nsubj_pobj_dobj\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 484,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Score: 88.577 {'Approach': 'nsubj_pobj_dobj', 'True Positive': 64, 'True Negative': 378, 'False Positive': 15, 'False Negative': 42}\n"
     ]
    }
   ],
   "source": [
    "# Print Variables\n",
    "print(\"Accuracy Score:\", accuracy_4, confusion_4)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Test Accuracy - Approach 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 485,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save Variables\n",
    "accuracy_5, confusion_5 = accuracy(\"nsubj_dobj\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 486,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Score: 84.569 {'Approach': 'nsubj_dobj', 'True Positive': 40, 'True Negative': 382, 'False Positive': 9, 'False Negative': 68}\n"
     ]
    }
   ],
   "source": [
    "# Print Variables\n",
    "print(\"Accuracy Score:\", accuracy_5, confusion_5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Test Accuracy - Approach 6"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 487,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save Variables\n",
    "accuracy_6, confusion_6 = accuracy(\"nsubj_poss\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 488,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Score: 84.569 {'Approach': 'nsubj_poss', 'True Positive': 43, 'True Negative': 379, 'False Positive': 16, 'False Negative': 61}\n"
     ]
    }
   ],
   "source": [
    "# Print Variables\n",
    "print(\"Accuracy Score:\", accuracy_6, confusion_6)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Test Accuracy - Approach 7"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 489,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save Variables\n",
    "accuracy_7, confusion_7 = accuracy(\"nsubj_pobj_dobj_poss\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 490,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Score: 88.577 {'Approach': 'nsubj_pobj_dobj_poss', 'True Positive': 68, 'True Negative': 374, 'False Positive': 26, 'False Negative': 31}\n"
     ]
    }
   ],
   "source": [
    "# Print Variables\n",
    "print(\"Accuracy Score:\", accuracy_7, confusion_7)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Test Accuracy - Approach 8"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 491,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save Variables\n",
    "accuracy_8, confusion_8 = accuracy(\"nsubj_pobj_poss\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 492,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Score: 88.778 {'Approach': 'nsubj_pobj_poss', 'True Positive': 67, 'True Negative': 376, 'False Positive': 22, 'False Negative': 34}\n"
     ]
    }
   ],
   "source": [
    "# Print Variables\n",
    "print(\"Accuracy Score:\", accuracy_8, confusion_8)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Test Accuracy - Approach 9"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 493,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save Variables\n",
    "accuracy_9, confusion_9 = accuracy(\"nsubj_dobj_poss\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 494,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Score: 84.369 {'Approach': 'nsubj_dobj_poss', 'True Positive': 44, 'True Negative': 377, 'False Positive': 21, 'False Negative': 57}\n"
     ]
    }
   ],
   "source": [
    "# Print Variables\n",
    "print(\"Accuracy Score:\", accuracy_9, confusion_9)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "> ### Test Accuracy - Approach 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 495,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save Variables\n",
    "accuracy_10, confusion_10 = accuracy(\"mention\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 496,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy Score: 79.96 {'Approach': 'mention', 'True Positive': 61, 'True Negative': 338, 'False Positive': 87, 'False Negative': 13}\n"
     ]
    }
   ],
   "source": [
    "# Print Variables\n",
    "print(\"Accuracy Score:\", accuracy_10, confusion_10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "------"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Visualising the results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Make a DataFrame from the results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 497,
   "metadata": {},
   "outputs": [],
   "source": [
    "results_df = pd.DataFrame([confusion_1, confusion_2, confusion_3, confusion_4, confusion_5, confusion_6, confusion_7, confusion_8, confusion_9, confusion_10])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 498,
   "metadata": {},
   "outputs": [],
   "source": [
    "results_df[\"Accuracy\"] = [accuracy_1, accuracy_2, accuracy_3, accuracy_4, accuracy_5, accuracy_6, accuracy_7, accuracy_8, accuracy_9, accuracy_10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 499,
   "metadata": {},
   "outputs": [
    {
     "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>Approach</th>\n",
       "      <th>False Negative</th>\n",
       "      <th>False Positive</th>\n",
       "      <th>True Negative</th>\n",
       "      <th>True Positive</th>\n",
       "      <th>Accuracy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>nsubj</td>\n",
       "      <td>72</td>\n",
       "      <td>3</td>\n",
       "      <td>384</td>\n",
       "      <td>40</td>\n",
       "      <td>84.970</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>nsubj_passsubj</td>\n",
       "      <td>72</td>\n",
       "      <td>3</td>\n",
       "      <td>384</td>\n",
       "      <td>40</td>\n",
       "      <td>84.970</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>nsubj_pobj</td>\n",
       "      <td>45</td>\n",
       "      <td>10</td>\n",
       "      <td>380</td>\n",
       "      <td>64</td>\n",
       "      <td>88.978</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>nsubj_pobj_dobj</td>\n",
       "      <td>42</td>\n",
       "      <td>15</td>\n",
       "      <td>378</td>\n",
       "      <td>64</td>\n",
       "      <td>88.577</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>nsubj_dobj</td>\n",
       "      <td>68</td>\n",
       "      <td>9</td>\n",
       "      <td>382</td>\n",
       "      <td>40</td>\n",
       "      <td>84.569</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>nsubj_poss</td>\n",
       "      <td>61</td>\n",
       "      <td>16</td>\n",
       "      <td>379</td>\n",
       "      <td>43</td>\n",
       "      <td>84.569</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>nsubj_pobj_dobj_poss</td>\n",
       "      <td>31</td>\n",
       "      <td>26</td>\n",
       "      <td>374</td>\n",
       "      <td>68</td>\n",
       "      <td>88.577</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>nsubj_pobj_poss</td>\n",
       "      <td>34</td>\n",
       "      <td>22</td>\n",
       "      <td>376</td>\n",
       "      <td>67</td>\n",
       "      <td>88.778</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>nsubj_dobj_poss</td>\n",
       "      <td>57</td>\n",
       "      <td>21</td>\n",
       "      <td>377</td>\n",
       "      <td>44</td>\n",
       "      <td>84.369</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>mention</td>\n",
       "      <td>13</td>\n",
       "      <td>87</td>\n",
       "      <td>338</td>\n",
       "      <td>61</td>\n",
       "      <td>79.960</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               Approach  False Negative  False Positive  True Negative  \\\n",
       "0  nsubj                 72              3               384             \n",
       "1  nsubj_passsubj        72              3               384             \n",
       "2  nsubj_pobj            45              10              380             \n",
       "3  nsubj_pobj_dobj       42              15              378             \n",
       "4  nsubj_dobj            68              9               382             \n",
       "5  nsubj_poss            61              16              379             \n",
       "6  nsubj_pobj_dobj_poss  31              26              374             \n",
       "7  nsubj_pobj_poss       34              22              376             \n",
       "8  nsubj_dobj_poss       57              21              377             \n",
       "9  mention               13              87              338             \n",
       "\n",
       "   True Positive  Accuracy  \n",
       "0  40             84.970    \n",
       "1  40             84.970    \n",
       "2  64             88.978    \n",
       "3  64             88.577    \n",
       "4  40             84.569    \n",
       "5  43             84.569    \n",
       "6  68             88.577    \n",
       "7  67             88.778    \n",
       "8  44             84.369    \n",
       "9  61             79.960    "
      ]
     },
     "execution_count": 499,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualise the Accuracy Scores"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 541,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 571,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1490.4x739.44 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Set Chart Options\n",
    "sns.set(rc={'figure.figsize':(20.7,10.27), })\n",
    "sns.set(style=\"darkgrid\")\n",
    "accuracy_chart = sns.barplot(x=\"Approach\", y=\"Accuracy\", data=results_df, color=\"#7CB7A3\")\n",
    "plt.ylim(70, 100)\n",
    "accuracy_chart.axes.set_title(\"Accuracy scores of target identification across 10 different approaches\",fontsize=30)\n",
    "accuracy_chart.set_xlabel(\"Approach\",fontsize=20)\n",
    "accuracy_chart.set_ylabel(\"Accuracy Score\",fontsize=20)\n",
    "accuracy_chart.tick_params(labelsize=12)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualise N False Positives"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 574,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1490.4x739.44 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "false_positives_chart = sns.barplot(x=\"Approach\", y=\"False Positive\", data=results_df, color=\"#B76E79\")\n",
    "plt.ylim(0, 100)\n",
    "false_positives_chart.axes.set_title(\"Number of false positives within target identification across 10 different approaches\",fontsize=30)\n",
    "false_positives_chart.set_xlabel(\"Approach\",fontsize=20)\n",
    "false_positives_chart.set_ylabel(\"Number of false positives\",fontsize=20)\n",
    "false_positives_chart.tick_params(labelsize=12)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Visualise Accuracy vs False Positives"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'plt' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-1-bb95f4c95bfd>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfigure\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfigsize\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m136\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m206\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0msns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrc\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0;34m'figure.figsize'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m20.7\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m10.27\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0maccuracy_vs_falsepositives\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlmplot\u001b[0m\u001b[0;34m(\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"Accuracy\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"False Positive\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mresults_df\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfit_reg\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhue\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Approach'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlegend\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mfalse_positives_chart\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0maxes\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_title\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Number of false positives within target identification across 10 different approaches\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mfontsize\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m30\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mfalse_positives_chart\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_xlabel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Approach\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mfontsize\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m20\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'plt' is not defined"
     ]
    }
   ],
   "source": [
    "plt.figure(figsize=(136, 206))\n",
    "sns.set(rc={'figure.figsize':(20.7,10.27), })\n",
    "accuracy_vs_falsepositives = sns.lmplot( x=\"Accuracy\", y=\"False Positive\", data=results_df, fit_reg=False, hue='Approach', legend=True)\n",
    "false_positives_chart.axes.set_title(\"Number of false positives within target identification across 10 different approaches\",fontsize=30)\n",
    "false_positives_chart.set_xlabel(\"Approach\",fontsize=20)\n",
    "false_positives_chart.set_ylabel(\"Number of false positives\",fontsize=20)\n",
    "false_positives_chart.tick_params(labelsize=12)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---- "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Where to go from here?\n",
    "> ##### Avenues I haven't explored:\n",
    "    - Co-reference Resolution: Identifying who 'He' is referring to \n",
    "    - Conditional adding of possessives / objects: Only adding target from a possessive reference if the possessive is part of the sentence subject \n",
    "    - Using a classifier to predict the target player\n",
    "    - Adding targets based on their position in an nltk parse tree\n",
    "    - Expanding annotated dataset to produce more detailed experiment"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}