{ "cells": [ { "cell_type": "markdown", "id": "4037071f", "metadata": {}, "source": [ "# fastTextによる分類\n", "\n", "[fastText](https://fasttext.cc/)\n", "は分類器の学習もサポートしていており、単語埋め込みも含めて学習することで精度の向上が期待できます。\n", "\n", "```{note}\n", "fastTextを使うためには事前にfasttextパッケージをインストールしておきます。\n", "\n", " !pip install fasttext==0.9.1\n", "```" ] }, { "cell_type": "markdown", "id": "c744ebe5", "metadata": {}, "source": [ "**データのロード**" ] }, { "cell_type": "code", "execution_count": 1, "id": "375b8518", "metadata": {}, "outputs": [], "source": [ "import pandas as pd\n", "from sklearn import model_selection\n", "\n", "data = pd.read_csv(\"input/pn_same_judge_preprocessed.csv\")\n", "train, test = model_selection.train_test_split(data, test_size=0.1, random_state=0)" ] }, { "cell_type": "markdown", "id": "6b82af17", "metadata": {}, "source": [ "## ラベルの準備\n", "\n", "fastTextが必要とするラベルを付与します。" ] }, { "cell_type": "code", "execution_count": 2, "id": "74bca0eb", "metadata": {}, "outputs": [], "source": [ "def apply_fn(row):\n", " tokens = row[\"tokens\"]\n", " label = f\"__label__{row['label_num']}\"\n", " return f\"{label} {tokens}\"" ] }, { "cell_type": "code", "execution_count": 3, "id": "6880dc0a", "metadata": {}, "outputs": [], "source": [ "for target in [train, test]:\n", " target[\"model_input\"] = target.apply(apply_fn, axis=\"columns\")" ] }, { "cell_type": "code", "execution_count": 4, "id": "881962e0", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
textlabel_numtokensmodel_input
2310また利用したいホテルである。0また 利用 する たい ホテル だ ある 。__label__0 また 利用 する たい ホテル だ ある 。
308お腹いっぱい食べてしまいました。0お腹 いっぱい 食べる て しまう ます た 。__label__0 お腹 いっぱい 食べる て しまう ます た 。
684とにかく狭い。1とにかく 狭い 。__label__1 とにかく 狭い 。
\n", "
" ], "text/plain": [ " text label_num tokens \\\n", "2310 また利用したいホテルである。 0 また 利用 する たい ホテル だ ある 。 \n", "308 お腹いっぱい食べてしまいました。 0 お腹 いっぱい 食べる て しまう ます た 。 \n", "684 とにかく狭い。 1 とにかく 狭い 。 \n", "\n", " model_input \n", "2310 __label__0 また 利用 する たい ホテル だ ある 。 \n", "308 __label__0 お腹 いっぱい 食べる て しまう ます た 。 \n", "684 __label__1 とにかく 狭い 。 " ] }, "execution_count": 4, "metadata": {}, "output_type": "execute_result" } ], "source": [ "train.head(n=3)" ] }, { "cell_type": "code", "execution_count": 5, "id": "363d0d21", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
textlabel_numtokensmodel_input
3574当日貸切状態で、最高の部屋を用意していただきました。0当日 貸切 状態 で 、 最高 の 部屋 を 用意 する て いただく ます た 。__label__0 当日 貸切 状態 で 、 最高 の 部屋 を 用意 する て いただく...
1386マタニティ旅行で利用しました。0マタニティ 旅行 で 利用 する ます た 。__label__0 マタニティ 旅行 で 利用 する ます た 。
499コンビニも近いです。0コンビニ も 近い です 。__label__0 コンビニ も 近い です 。
\n", "
" ], "text/plain": [ " text label_num \\\n", "3574 当日貸切状態で、最高の部屋を用意していただきました。 0 \n", "1386 マタニティ旅行で利用しました。 0 \n", "499 コンビニも近いです。 0 \n", "\n", " tokens \\\n", "3574 当日 貸切 状態 で 、 最高 の 部屋 を 用意 する て いただく ます た 。 \n", "1386 マタニティ 旅行 で 利用 する ます た 。 \n", "499 コンビニ も 近い です 。 \n", "\n", " model_input \n", "3574 __label__0 当日 貸切 状態 で 、 最高 の 部屋 を 用意 する て いただく... \n", "1386 __label__0 マタニティ 旅行 で 利用 する ます た 。 \n", "499 __label__0 コンビニ も 近い です 。 " ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "test.head(n=3)" ] }, { "cell_type": "markdown", "id": "7257055b", "metadata": {}, "source": [ "ファイルに保存します。" ] }, { "cell_type": "code", "execution_count": 6, "id": "927755a1", "metadata": {}, "outputs": [], "source": [ "train[[\"model_input\"]].to_csv(\"input/pn_ft_train.csv\", header=None, index=None)\n", "test[[\"model_input\"]].to_csv(\"input/pn_ft_test.csv\", header=None, index=None)" ] }, { "cell_type": "code", "execution_count": 7, "id": "127ce02c", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "__label__0 また 利用 する たい ホテル だ ある 。\r\n", "__label__0 お腹 いっぱい 食べる て しまう ます た 。\r\n", "__label__1 とにかく 狭い 。\r\n" ] } ], "source": [ "!head -n3 input/pn_ft_train.csv" ] }, { "cell_type": "code", "execution_count": 8, "id": "f5c477ec", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "__label__0 当日 貸切 状態 で 、 最高 の 部屋 を 用意 する て いただく ます た 。\r\n", "__label__0 マタニティ 旅行 で 利用 する ます た 。\r\n", "__label__0 コンビニ も 近い です 。\r\n" ] } ], "source": [ "!head -n3 input/pn_ft_test.csv" ] }, { "cell_type": "markdown", "id": "2e8f850c", "metadata": {}, "source": [ "## 学習する" ] }, { "cell_type": "code", "execution_count": 9, "id": "449e63f7", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "Read 0M words\n", "Number of words: 3075\n", "Number of labels: 2\n", "Progress: 100.0% words/sec/thread: 1201377 lr: 0.000000 loss: 0.212663 ETA: 0h 0m\n" ] } ], "source": [ "import fasttext\n", "\n", "model = fasttext.train_supervised(input=\"input/pn_ft_train.csv\")" ] }, { "cell_type": "code", "execution_count": 10, "id": "329839b8", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "(419, 0.9093078758949881, 0.9093078758949881)" ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" } ], "source": [ "model.test(\"input/pn_ft_test.csv\")" ] }, { "cell_type": "code", "execution_count": 11, "id": "b01932cf", "metadata": {}, "outputs": [], "source": [ "def prob_fn(item):\n", " input_text = \" \".join(item.split()[1:])\n", " pred = model.predict(input_text)\n", " label = pred[0][0]\n", " score = pred[1][0]\n", " if label == \"__label__1\":\n", " pass\n", " elif label == \"__label__0\":\n", " score = 1 - score\n", " else:\n", " raise Exception(f\"Label is not expected one: {label}\")\n", " label_map = {\"__label__1\": 1, \"__label__0\": 0}\n", " return label_map[label], score\n", "\n", "pred = test[\"model_input\"].apply(lambda x: prob_fn(x)[0])\n", "score = test[\"model_input\"].apply(lambda x: prob_fn(x)[1])" ] }, { "cell_type": "code", "execution_count": 12, "id": "0635ad61", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "3574 0\n", "1386 0\n", "499 0\n", "3756 0\n", "914 0\n", "Name: model_input, dtype: int64" ] }, "execution_count": 12, "metadata": {}, "output_type": "execute_result" } ], "source": [ "pred.head()" ] }, { "cell_type": "code", "execution_count": 13, "id": "439d1270", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "3574 0.017346\n", "1386 0.000513\n", "499 0.006268\n", "3756 0.018718\n", "914 0.019504\n", "Name: model_input, dtype: float64" ] }, "execution_count": 13, "metadata": {}, "output_type": "execute_result" } ], "source": [ "score.head()" ] }, { "cell_type": "code", "execution_count": 14, "id": "595a277c", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 14, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "from sklearn.metrics import ConfusionMatrixDisplay\n", "\n", "ConfusionMatrixDisplay.from_predictions(y_true=test[\"label_num\"], y_pred=pred)" ] }, { "cell_type": "code", "execution_count": 15, "id": "deb451ec", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "" ] }, "execution_count": 15, "metadata": {}, "output_type": "execute_result" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "from sklearn.metrics import PrecisionRecallDisplay\n", "\n", "PrecisionRecallDisplay.from_predictions(y_true=test[\"label_num\"], y_pred=score, name=\"fastText\")" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.8.6" } }, "nbformat": 4, "nbformat_minor": 5 }