{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Using BERTopic" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "
\n", "\n", "This tutorial is available as an IPython notebook at [Malaya/example/topic-modeling-bertopic](https://github.com/huseinzol05/Malaya/tree/master/example/topic-modeling-bertopic).\n", " \n", "
" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Install BERTopic" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "Make sure you already installed `BERTopic`,\n", "\n", "```bash\n", "pip3 install bertopic\n", "```" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "You can read more about BERTopic at https://maartengr.github.io/BERTopic/#quick-start" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "import pandas as pd\n", "import malaya" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "df = pd.read_csv('tests/02032018.csv',sep=';')\n", "df = df.iloc[3:,1:]\n", "df.columns = ['text','label']\n", "corpus = df.text.tolist()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Load vectorizer model" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "(2, 256)" ] }, "execution_count": 4, "metadata": {}, "output_type": "execute_result" } ], "source": [ "vectorizer = malaya.transformer.huggingface()\n", "vectorizer.vectorize(['hello', 'ayam']).shape" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Create embedder object" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [], "source": [ "from bertopic.backend import BaseEmbedder\n", "import numpy as np\n", "from typing import List" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [], "source": [ "class embedder(BaseEmbedder):\n", "\n", " def embed(self,\n", " documents: List[str],\n", " verbose: bool = False):\n", " return vectorizer.vectorize(documents)\n", "\n", " def embed_words(self,\n", " words: List[str],\n", " verbose: bool = False):\n", " return self.embed(words, verbose)\n", "\n", " def embed_documents(self,\n", " document: List[str],\n", " verbose: bool = False):\n", " return self.embed(document, verbose)" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [], "source": [ "embedder = embedder()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Train BERTopic" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [], "source": [ "from bertopic import BERTopic\n", "from bertopic.backend._utils import select_backend\n", "\n", "model = BERTopic(language = None, embedding_model = embedder)\n", "model.language" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [], "source": [ "embedding_model = select_backend(model.embedding_model,\n", " language=model.language)" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [], "source": [ "topics, probs = model.fit_transform(corpus)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Get topic frequencies" ] }, { "cell_type": "code", "execution_count": 21, "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", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
TopicCount
0-191
1074
2150
3232
4321
5418
6518
\n", "
" ], "text/plain": [ " Topic Count\n", "0 -1 91\n", "1 0 74\n", "2 1 50\n", "3 2 32\n", "4 3 21\n", "5 4 18\n", "6 5 18" ] }, "execution_count": 21, "metadata": {}, "output_type": "execute_result" } ], "source": [ "model.get_topic_freq()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Get topic" ] }, { "cell_type": "code", "execution_count": 25, "metadata": { "scrolled": true }, "outputs": [ { "data": { "text/plain": [ "[('dan', 0.06565089788198161),\n", " ('yang', 0.05259762576836089),\n", " ('untuk', 0.040200581210431734),\n", " ('dalam', 0.03976023208988511),\n", " ('ini', 0.03830356403041938),\n", " ('malaysia', 0.03714433613321695),\n", " ('akan', 0.03629905145154332),\n", " ('lebih', 0.03627484217552929),\n", " ('kami', 0.035442639029197864),\n", " ('di', 0.03322709631876396)]" ] }, "execution_count": 25, "metadata": {}, "output_type": "execute_result" } ], "source": [ "model.get_topic(0)[:10]" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### Find similar topics" ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "[1, 5, 2, 3, 4, 0, -1]" ] }, "execution_count": 20, "metadata": {}, "output_type": "execute_result" } ], "source": [ "similar_topics, similarity = model.find_topics('najib', top_n=10)\n", "similar_topics" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.8.10" } }, "nbformat": 4, "nbformat_minor": 4 }