classifying_works_of_fiction.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Author** Andrew Reece  \n",
    "**Contact** andrew.reece@gmail.com &lt; <a href=\"https://www.linkedin.com/in/andrewreece/\">li</a> &gt; &lt; <a href=\"https://stackoverflow.com/users/2799941/andrew-reece\">so</a> &gt;  \n",
    "**Date** 23 November 2018"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Motivation**  \n",
    "Can we train a language model to distinguish between classical works of fiction?  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "from fastai import *\n",
    "from fastai.text import *\n",
    "from fastai.vision import ClassificationInterpretation\n",
    "import pandas as pd\n",
    "import pathlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "%reload_ext autoreload\n",
    "%autoreload 2\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Here is a subset of works of fiction from Project Gutenberg (in .tar.gz format):    \n",
    "https://drive.google.com/open?id=1Atth50tE_J7FGRBwsr2Mr11DC0i1JBRQ\n",
    "\n",
    "Untar it in the same directory as this notebook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#! wget https://drive.google.com/open?id=1Atth50tE_J7FGRBwsr2Mr11DC0i1JBRQ\n",
    "#! tar -zxvf gutenberg.tar.gz"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "gut_path = pathlib.Path(\"gutenberg/\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Fiction from Project Gutenberg\n",
    "\n",
    "The Gutenberg metadata has been removed from these files, and the first line gives the title, author, and publication year in a systematic pattern.  \n",
    "In total, this dataset contains 26 works of fiction."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "26"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gutenberg_filenames = list(gut_path.glob(\"*.txt\"))\n",
    "len(gutenberg_filenames)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "File names follow the convention `<author_abbreviation>-<title_abbreviation>`.  \n",
    "You can isolate the file name with the following command"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'twain-huckleberry_finn'"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gutenberg_filenames[0].name.split(\".\")[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Preprocessing**  \n",
    "The function below takes a simple approach to splitting up a text by paragraphs.  \n",
    "We'll use the paragraphs created from `gutenberg_iterator` as inputs to our language model.  \n",
    "\n",
    "This code is a modified version of a function originally written by <a href=\"https://web.stanford.edu/~cgpotts/\">Chris Potts</a>, for <a href=\"http://web.stanford.edu/class/linguist278/\">Linguistics 278</a> at Stanford. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "def gutenberg_iterator(filename):\n",
    "    \"\"\"Yields paragraphs (as defined simply by multiple \n",
    "    newlines in a row).\n",
    "    \n",
    "    Parameters\n",
    "    ----------\n",
    "    filename : str\n",
    "        Full path to the file.\n",
    "        \n",
    "    Yields\n",
    "    ------\n",
    "    multiline str\n",
    "    \n",
    "    \"\"\"\n",
    "    with open(filename) as f:\n",
    "        contents = f.read()\n",
    "    for para in re.split(r\"[\\n\\s*]{2,}\", contents):\n",
    "        try:\n",
    "            if para[0] != \"[\" and not para.split()[0].isupper():\n",
    "                yield para\n",
    "        except IndexError:\n",
    "            continue"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Split up each text into train/validation data frames. Build two master data frames of train and validation data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_frac = .7\n",
    "\n",
    "train_master = pd.DataFrame()\n",
    "valid_master = pd.DataFrame()\n",
    "\n",
    "for i, fn in enumerate(gutenberg_filenames):\n",
    "    paras = list()\n",
    "    para_iterator = gutenberg_iterator(fn)\n",
    "    for para in para_iterator:\n",
    "        paras.append([i, para])\n",
    "    paras_df = pd.DataFrame(paras, columns=[\"label\", \"text\"])\n",
    "    train_ix = paras_df.sample(frac=train_frac).index\n",
    "    train_mask = paras_df.index.isin(train_ix)\n",
    "    train_master = pd.concat([train_master, paras_df.loc[train_mask]])\n",
    "    valid_master = pd.concat([valid_master, paras_df.iloc[~train_mask]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_master.to_csv(\"gutenberg_train.csv\", index=False)\n",
    "valid_master.to_csv(\"gutenberg_valid.csv\", index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Save a legend of filename-index pairs.  \n",
    "We'll use the serial integer indexes for the label column in the data frames we pass into `Learner`, but we'll want to replace those with readable labels, later on."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.Series([fn.name.split(\".\")[0] for fn in gutenberg_filenames], name=\"text\")\\\n",
    "  .reset_index()\\\n",
    "  .to_csv(\"label_key.csv\", index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load `train` and `valid` data frames."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "train = pd.read_csv(\"gutenberg_train.csv\")\n",
    "valid = pd.read_csv(\"gutenberg_valid.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We have a bit of a class imbalance issue, which will make the use of a raw accuracy metric somewhat misleading.  \n",
    "We could use the `f_beta` metric instead, although this current notebook hasn't incorporated that yet. For now, we can just keep in mind which are the over-/under-represented classes, and eyeball the resulting confusion matrix to see if predictive accuracy is skewed or tends to vote for majority classes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3     3797\n",
       "11    3272\n",
       "10    2411\n",
       "21    2330\n",
       "24    2279\n",
       "23    2141\n",
       "15    1657\n",
       "14    1541\n",
       "25    1461\n",
       "0     1415\n",
       "12    1209\n",
       "16    1165\n",
       "18     972\n",
       "4      758\n",
       "2      752\n",
       "5      523\n",
       "6      388\n",
       "19     362\n",
       "13     338\n",
       "7      317\n",
       "9      271\n",
       "1      270\n",
       "20     256\n",
       "8      241\n",
       "22      67\n",
       "17      57\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "valid.label.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Build a language model across all Gutenberg texts\n",
    "\n",
    "This section largely follows `lesson3-imdb` from the fast.ai Fall 2018 DL1 course.  \n",
    "One difference is that it uses the `from_df()` factory method when creating a `DataBunch`.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = pathlib.Path(\".\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_lm = TextDataBunch.from_df(path, train_df=train, valid_df=valid, text_cols=\"text\", label_cols=\"label\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_lm.save(\"tmp_lm\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['xxunk', 'xxpad', ',', '\\n', 'the', '1', 'xxbos', 'xxfld', '.', 'and']"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_lm.vocab.itos[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text xxbos xxfld 1 the missouri negro \n",
       " dialect ; the extremest form of the backwoods xxunk dialect ; the \n",
       " ordinary \" pike county \" dialect ; and four modified varieties of this last . \n",
       " the xxunk have not been done in a haphazard fashion , or by xxunk ; \n",
       " but xxunk , and with the trustworthy guidance and support of \n",
       " personal familiarity with these several forms of speech ."
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_lm.train_ds[0][0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "bs = 48"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "# I'm not sure we need a TextDataBunch first and then a TextLMDataBunch - this is just following the Lesson 3 code.\n",
    "data_lm = TextLMDataBunch.load(path, 'tmp_lm', bs = bs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = language_model_learner(data_lm, pretrained_model=URLs.WT103, drop_mult=0.3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n"
     ]
    }
   ],
   "source": [
    "learn.lr_find()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.recorder.plot(skip_end=15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total time: 01:18\n",
      "epoch  train_loss  valid_loss  accuracy\n",
      "1      5.316948    5.060972    0.212872  (01:18)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "learn.fit_one_cycle(1, 3e-1, moms=(0.8,0.7))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn.save('fit_head')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total time: 15:07\n",
      "epoch  train_loss  valid_loss  accuracy\n",
      "1      4.725402    4.617736    0.260116  (01:30)\n",
      "2      4.486015    4.399654    0.278521  (01:30)\n",
      "3      4.359496    4.273832    0.287831  (01:30)\n",
      "4      4.216336    4.164670    0.294202  (01:31)\n",
      "5      4.098336    4.081409    0.302391  (01:30)\n",
      "6      3.995981    4.029743    0.308258  (01:30)\n",
      "7      3.881822    3.990638    0.312243  (01:30)\n",
      "8      3.756205    3.983469    0.313753  (01:30)\n",
      "9      3.662281    3.987786    0.314340  (01:30)\n",
      "10     3.613561    3.998944    0.314088  (01:30)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "learn.unfreeze()\n",
    "learn.fit_one_cycle(10, 1e-2, moms=(0.8,0.7))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn.save_encoder('fine_tuned_enc')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create text classifier using trained encoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<table>  <col width='90%'>  <col width='10%'>  <tr>\n",
       "    <th>text</th>\n",
       "    <th>label</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>xxbos xxfld 1 now he lapsed into suffering again , as the dry argument was resumed . \\n presently he bethought him of a treasure he had and got it out . it was \\n a large black beetle with formidable jaws -- a \" pinchbug , \" he called</th>\n",
       "    <th>5</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>xxbos xxfld 1 circumstances that might swell to half an hour 's relation , \\n and contained multiplied proofs to her who had seen them , had passed \\n xxunk by her who now heard them ; but the two latest occurrences \\n to be mentioned , the two of</th>\n",
       "    <th>25</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>xxbos xxfld 1 \" hang the boy , ca n't i never learn anything ? ai n't he played me tricks \\n enough like that for me to be looking out for him by this time ? but old \\n fools is the biggest fools there is . ca n't</th>\n",
       "    <th>5</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>xxbos xxfld 1 mur . wherefore reioyce ? \\n what conquest brings he home ? \\n what xxunk follow him to rome , \\n to grace in xxunk bonds his chariot xxunk ? \\n you xxunk , you stones , you worse then xxunk things : \\n o you hard</th>\n",
       "    <th>9</th>\n",
       "  </tr>\n",
       "  <tr>\n",
       "    <th>xxbos xxfld 1 her brother 's return was the first comfort ; he could take best care \\n of his wife ; and the second blessing was the arrival of the apothecary . \\n till he came and had examined the child , their apprehensions were \\n the worse for</th>\n",
       "    <th>4</th>\n",
       "  </tr>\n",
       "</table>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data_clas = TextClasDataBunch.load(path)\n",
    "data_clas.show_batch()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = text_classifier_learner(data_clas, drop_mult=0.5)\n",
    "learn.load_encoder('fine_tuned_enc')\n",
    "learn.freeze()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LR Finder is complete, type {learner_name}.recorder.plot() to see the graph.\n"
     ]
    }
   ],
   "source": [
    "learn.lr_find()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "learn.recorder.plot()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total time: 00:49\n",
      "epoch  train_loss  valid_loss  accuracy\n",
      "1      1.669107    1.360316    0.554678  (00:49)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "learn.fit_one_cycle(1, 5e-2, moms=(0.8,0.7))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn.save('first_gut')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total time: 01:02\n",
      "epoch  train_loss  valid_loss  accuracy\n",
      "1      1.247227    1.038880    0.664926  (01:02)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "learn.freeze_to(-2)\n",
    "learn.fit_one_cycle(1, slice(1e-2/(2.6**4),1e-2), moms=(0.8,0.7))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn.save('second_gut')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total time: 01:35\n",
      "epoch  train_loss  valid_loss  accuracy\n",
      "1      1.097242    0.934629    0.694545  (01:35)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "learn.freeze_to(-3)\n",
    "learn.fit_one_cycle(1, slice(5e-3/(2.6**4),5e-3), moms=(0.8,0.7))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn.save('third_gut')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total time: 04:22\n",
      "epoch  train_loss  valid_loss  accuracy\n",
      "1      1.026059    0.902645    0.703074  (02:09)\n",
      "2      0.996181    0.905725    0.707372  (02:12)\n",
      "\n"
     ]
    }
   ],
   "source": [
    "learn.unfreeze()\n",
    "learn.fit_one_cycle(2, slice(1e-3/(2.6**4),1e-3), moms=(0.8,0.7))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With just a bit of training, we achieve 71% classification accuracy over 26 classes."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Understanding our model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can look at classification accuracy using the `ClassificationInterpretation` class from the `fastai.vision` module."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [],
   "source": [
    "interp = ClassificationInterpretation.from_learner(learn)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Load the index-text legend.  This will allow us to replace the numeric classes with more readable text labels."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "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>index</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>twain-huckleberry_finn</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>blake-poems</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>chesterton-brown</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>dickens-ncklb10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>austen-persuasion</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   index                    text\n",
       "0      0  twain-huckleberry_finn\n",
       "1      1             blake-poems\n",
       "2      2        chesterton-brown\n",
       "3      3         dickens-ncklb10\n",
       "4      4       austen-persuasion"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label_key = pd.read_csv(\"label_key.csv\")\n",
    "label_key_dict = {x[\"index\"]: x[\"text\"] for x in label_key.to_dict(orient=\"records\")}\n",
    "\n",
    "label_key.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "confusion = interp.confusion_matrix()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that `plot_confusion_matrix()` creates a `plt` object, which can be adjusted simply by referring to `plt` afterwards.  Here, we use this knowledge to replace the tick labels (numeric indices) on the confusion matrix with text labels."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 840x840 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "interp.plot_confusion_matrix(figsize=(14,14), dpi=60)\n",
    "_ = plt.xticks(np.arange(26), label_key.text.values)\n",
    "_ = plt.yticks(np.arange(26), label_key.text.values)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Recall the class imbalance we discovered earlier - it looks like even for classes with very low sample size, accuracy is pretty good and it's not terribly skewed to majority class predictions."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The `most_confused` method gives a clearer sense for where the model goes wrong.  The most frequent misses are, unsurprisingly, when one author has multiple texts and the model predicts the right author, but the wrong book. More interestingly, perhaps, is that authors with similar writing styles (Blake and Milton, for example) are confused, whereas authors with very different styles (Lewis Carroll vs Charles Dickens) are rarely confused.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "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>Actual</th>\n",
       "      <th>Predicted</th>\n",
       "      <th>n_errors</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>dickens-ncklb10</td>\n",
       "      <td>dickens-olivr11</td>\n",
       "      <td>389</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>dickens-olivr11</td>\n",
       "      <td>dickens-ncklb10</td>\n",
       "      <td>389</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>christie-masac11</td>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>180</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>christie-secad10</td>\n",
       "      <td>christie-masac11</td>\n",
       "      <td>167</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>edgeworth-parents</td>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>162</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>blake-poems</td>\n",
       "      <td>milton-paradise</td>\n",
       "      <td>158</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>christie-masac11</td>\n",
       "      <td>christie-secad10</td>\n",
       "      <td>157</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>austen-persuasion</td>\n",
       "      <td>austen-emma</td>\n",
       "      <td>129</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>christie-masac11</td>\n",
       "      <td>128</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>melville-moby_dick</td>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>125</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>dickens-ncklb10</td>\n",
       "      <td>edgeworth-parents</td>\n",
       "      <td>115</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>dickens-ncklb10</td>\n",
       "      <td>melville-moby_dick</td>\n",
       "      <td>113</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>edgeworth-parents</td>\n",
       "      <td>112</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>dickens-ncklb10</td>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>111</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>melville-moby_dick</td>\n",
       "      <td>edgeworth-parents</td>\n",
       "      <td>101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>shakespeare-macbeth</td>\n",
       "      <td>shakespeare-caesar</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>edgeworth-parents</td>\n",
       "      <td>melville-moby_dick</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>christie-secad10</td>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>whitman-leaves</td>\n",
       "      <td>milton-paradise</td>\n",
       "      <td>95</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>chesterton-brown</td>\n",
       "      <td>chesterton-ball</td>\n",
       "      <td>94</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>chesterton-ball</td>\n",
       "      <td>melville-moby_dick</td>\n",
       "      <td>94</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>austen-sense</td>\n",
       "      <td>austen-emma</td>\n",
       "      <td>90</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>christie-secad10</td>\n",
       "      <td>88</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>melville-moby_dick</td>\n",
       "      <td>86</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>twain-huckleberry_finn</td>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>85</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>chesterton-ball</td>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>85</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>edgeworth-parents</td>\n",
       "      <td>christie-secad10</td>\n",
       "      <td>83</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>carroll-looking-glass</td>\n",
       "      <td>carroll-alice</td>\n",
       "      <td>82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>twain-huckleberry_finn</td>\n",
       "      <td>christie-secad10</td>\n",
       "      <td>81</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>austen-emma</td>\n",
       "      <td>austen-sense</td>\n",
       "      <td>79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>244</th>\n",
       "      <td>whitman-leaves</td>\n",
       "      <td>dickens-ncklb10</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>245</th>\n",
       "      <td>whitman-leaves</td>\n",
       "      <td>christie-masac11</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>246</th>\n",
       "      <td>blake-songs</td>\n",
       "      <td>dickens-olivr11</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>247</th>\n",
       "      <td>austen-sense</td>\n",
       "      <td>twain-huckleberry_finn</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>248</th>\n",
       "      <td>austen-sense</td>\n",
       "      <td>chesterton-brown</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>249</th>\n",
       "      <td>austen-sense</td>\n",
       "      <td>melville-moby_dick</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>250</th>\n",
       "      <td>shakespeare-hamlet</td>\n",
       "      <td>whitman-leaves</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>251</th>\n",
       "      <td>burgess-busterbrown</td>\n",
       "      <td>twain-tom_sawyer</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>252</th>\n",
       "      <td>dickens-olivr11</td>\n",
       "      <td>austen-sense</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>253</th>\n",
       "      <td>dickens-olivr11</td>\n",
       "      <td>carroll-alice</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>254</th>\n",
       "      <td>austen-emma</td>\n",
       "      <td>twain-tom_sawyer</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>255</th>\n",
       "      <td>twain-huckleberry_finn</td>\n",
       "      <td>chesterton-brown</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>256</th>\n",
       "      <td>twain-huckleberry_finn</td>\n",
       "      <td>milton-paradise</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>257</th>\n",
       "      <td>twain-huckleberry_finn</td>\n",
       "      <td>whitman-leaves</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>258</th>\n",
       "      <td>blake-poems</td>\n",
       "      <td>chesterton-ball</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>259</th>\n",
       "      <td>austen-persuasion</td>\n",
       "      <td>dickens-olivr11</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>260</th>\n",
       "      <td>bryant-stories</td>\n",
       "      <td>christie-masac11</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>261</th>\n",
       "      <td>carroll-looking-glass</td>\n",
       "      <td>twain-huckleberry_finn</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>262</th>\n",
       "      <td>carroll-looking-glass</td>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>263</th>\n",
       "      <td>edgeworth-parents</td>\n",
       "      <td>shakespeare-hamlet</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>264</th>\n",
       "      <td>whitman-leaves</td>\n",
       "      <td>austen-sense</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>265</th>\n",
       "      <td>chesterton-ball</td>\n",
       "      <td>austen-persuasion</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>266</th>\n",
       "      <td>chesterton-ball</td>\n",
       "      <td>twain-tom_sawyer</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>267</th>\n",
       "      <td>chesterton-ball</td>\n",
       "      <td>shakespeare-caesar</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>268</th>\n",
       "      <td>dickens-grexp10</td>\n",
       "      <td>austen-persuasion</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>269</th>\n",
       "      <td>burgess-busterbrown</td>\n",
       "      <td>milton-paradise</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>270</th>\n",
       "      <td>melville-moby_dick</td>\n",
       "      <td>bryant-stories</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>271</th>\n",
       "      <td>melville-moby_dick</td>\n",
       "      <td>shakespeare-macbeth</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>272</th>\n",
       "      <td>melville-moby_dick</td>\n",
       "      <td>blake-songs</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>273</th>\n",
       "      <td>melville-moby_dick</td>\n",
       "      <td>austen-sense</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>274 rows × 3 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                     Actual               Predicted  n_errors\n",
       "0           dickens-ncklb10         dickens-olivr11       389\n",
       "1           dickens-olivr11         dickens-ncklb10       389\n",
       "2          christie-masac11         dickens-grexp10       180\n",
       "3          christie-secad10        christie-masac11       167\n",
       "4         edgeworth-parents         dickens-grexp10       162\n",
       "5               blake-poems         milton-paradise       158\n",
       "6          christie-masac11        christie-secad10       157\n",
       "7         austen-persuasion             austen-emma       129\n",
       "8           dickens-grexp10        christie-masac11       128\n",
       "9        melville-moby_dick         dickens-grexp10       125\n",
       "10          dickens-ncklb10       edgeworth-parents       115\n",
       "11          dickens-ncklb10      melville-moby_dick       113\n",
       "12          dickens-grexp10       edgeworth-parents       112\n",
       "13          dickens-ncklb10         dickens-grexp10       111\n",
       "14       melville-moby_dick       edgeworth-parents       101\n",
       "15      shakespeare-macbeth      shakespeare-caesar       100\n",
       "16        edgeworth-parents      melville-moby_dick       100\n",
       "17         christie-secad10         dickens-grexp10       100\n",
       "18           whitman-leaves         milton-paradise        95\n",
       "19         chesterton-brown         chesterton-ball        94\n",
       "20          chesterton-ball      melville-moby_dick        94\n",
       "21             austen-sense             austen-emma        90\n",
       "22          dickens-grexp10        christie-secad10        88\n",
       "23          dickens-grexp10      melville-moby_dick        86\n",
       "24   twain-huckleberry_finn         dickens-grexp10        85\n",
       "25          chesterton-ball         dickens-grexp10        85\n",
       "26        edgeworth-parents        christie-secad10        83\n",
       "27    carroll-looking-glass           carroll-alice        82\n",
       "28   twain-huckleberry_finn        christie-secad10        81\n",
       "29              austen-emma            austen-sense        79\n",
       "..                      ...                     ...       ...\n",
       "244          whitman-leaves         dickens-ncklb10         4\n",
       "245          whitman-leaves        christie-masac11         4\n",
       "246             blake-songs         dickens-olivr11         4\n",
       "247            austen-sense  twain-huckleberry_finn         4\n",
       "248            austen-sense        chesterton-brown         4\n",
       "249            austen-sense      melville-moby_dick         4\n",
       "250      shakespeare-hamlet          whitman-leaves         4\n",
       "251     burgess-busterbrown        twain-tom_sawyer         4\n",
       "252         dickens-olivr11            austen-sense         4\n",
       "253         dickens-olivr11           carroll-alice         4\n",
       "254             austen-emma        twain-tom_sawyer         4\n",
       "255  twain-huckleberry_finn        chesterton-brown         3\n",
       "256  twain-huckleberry_finn         milton-paradise         3\n",
       "257  twain-huckleberry_finn          whitman-leaves         3\n",
       "258             blake-poems         chesterton-ball         3\n",
       "259       austen-persuasion         dickens-olivr11         3\n",
       "260          bryant-stories        christie-masac11         3\n",
       "261   carroll-looking-glass  twain-huckleberry_finn         3\n",
       "262   carroll-looking-glass         dickens-grexp10         3\n",
       "263       edgeworth-parents      shakespeare-hamlet         3\n",
       "264          whitman-leaves            austen-sense         3\n",
       "265         chesterton-ball       austen-persuasion         3\n",
       "266         chesterton-ball        twain-tom_sawyer         3\n",
       "267         chesterton-ball      shakespeare-caesar         3\n",
       "268         dickens-grexp10       austen-persuasion         3\n",
       "269     burgess-busterbrown         milton-paradise         3\n",
       "270      melville-moby_dick          bryant-stories         3\n",
       "271      melville-moby_dick     shakespeare-macbeth         3\n",
       "272      melville-moby_dick             blake-songs         3\n",
       "273      melville-moby_dick            austen-sense         3\n",
       "\n",
       "[274 rows x 3 columns]"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labeled_classes = [(label_key_dict[x[0]], label_key_dict[x[1]], x[2]) \n",
    "                   for x in interp.most_confused(min_val=2)]\n",
    "pd.DataFrame(labeled_classes, columns=[\"Actual\", \"Predicted\", \"n_errors\"])"
   ]
  }
 ],
 "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": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}