2016-05-26 12:33:27 +00:00
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"![](images/EscUpmPolit_p.gif \"UPM\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Course Notes for Learning Intelligent Systems"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
2019-03-06 16:46:12 +00:00
"Department of Telematic Engineering Systems, Universidad Politécnica de Madrid, © Carlos A. Iglesias"
2016-05-26 12:33:27 +00:00
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Semantic Models"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Table of Contents\n",
"* [Objectives](#Objectives)\n",
"* [Corpus](#Corpus)\n",
"* [Converting Scikit-learn to gensim](#Converting-Scikit-learn-to-gensim)\n",
"* [Latent Dirichlet Allocation (LDA)](#Latent-Dirichlet-Allocation-%28LDA%29)\n",
"* [Latent Semantic Indexing (LSI)](#Latent-Semantic-Indexing-%28LSI%29)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Objectives"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"In this session we provide a quick overview of the semantic models presented during the classes. In this case, we will use a real corpus so that we can extract meaningful patterns.\n",
"\n",
"The main objectives of this session are:\n",
"* Understand the models and their differences\n",
"* Learn to use some of the most popular NLP libraries"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Corpus"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We are going to use on of the corpus that come prepackaged with Scikit-learn: the [20 newsgroup datase](http://qwone.com/~jason/20Newsgroups/). The 20 newsgroup dataset contains 20k documents that belong to 20 topics.\n",
"\n",
"We inspect now the corpus using the facilities from Scikit-learn, as explain in [scikit-learn](http://scikit-learn.org/stable/datasets/twenty_newsgroups.html#newsgroups)"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 1,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(2034, 2807)"
]
},
"execution_count": 1,
"metadata": {},
"output_type": "execute_result"
}
],
2016-05-26 12:33:27 +00:00
"source": [
"from sklearn.datasets import fetch_20newsgroups\n",
"\n",
"# We filter only some categories, otherwise we have 20 categories\n",
"categories = ['alt.atheism', 'talk.religion.misc', 'comp.graphics', 'sci.space']\n",
"# We remove metadata to avoid bias in the classification\n",
"newsgroups_train = fetch_20newsgroups(subset='train', \n",
" remove=('headers', 'footers', 'quotes'), \n",
" categories=categories)\n",
"newsgroups_test = fetch_20newsgroups(subset='test', remove=('headers', 'footers', 'quotes'),\n",
" categories=categories)\n",
"\n",
"\n",
"# Obtain a vector\n",
"\n",
"from sklearn.feature_extraction.text import TfidfVectorizer\n",
"\n",
"vectorizer = TfidfVectorizer(analyzer='word', stop_words='english', min_df=10)\n",
"\n",
"vectors_train = vectorizer.fit_transform(newsgroups_train.data)\n",
"vectors_train.shape"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Converting Scikit-learn to gensim"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Although scikit-learn provides an LDA implementation, it is more popular the package *gensim*, which also provides an LSI implementation, as well as other functionalities. Fortunately, scikit-learn sparse matrices can be used in Gensim using the function *matutils.Sparse2Corpus()*. Anyway, if you are using intensively LDA,it can be convenient to create the corpus with their functions.\n",
"\n",
"You should install first *gensim*. Run 'conda install -c anaconda gensim=0.12.4' in a terminal."
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 2,
2019-02-28 14:30:33 +00:00
"metadata": {},
2016-05-26 12:33:27 +00:00
"outputs": [],
"source": [
"from gensim import matutils\n",
"\n",
"vocab = vectorizer.get_feature_names()\n",
"\n",
"dictionary = dict([(i, s) for i, s in enumerate(vectorizer.get_feature_names())])\n",
"corpus_tfidf = matutils.Sparse2Corpus(vectors_train)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Latent Dirichlet Allocation (LDA)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Although scikit-learn provides an LDA implementation, it is more popular the package *gensim*, which also provides an LSI implementation, as well as other functionalities. Fortunately, scikit-learn sparse matrices can be used in Gensim using the function *matutils.Sparse2Corpus()*."
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 3,
2019-02-28 14:30:33 +00:00
"metadata": {},
2016-05-26 12:33:27 +00:00
"outputs": [],
"source": [
"from gensim.models.ldamodel import LdaModel\n",
"\n",
"# It takes a long time\n",
"\n",
"# train the lda model, choosing number of topics equal to 4\n",
"lda = LdaModel(corpus_tfidf, num_topics=4, passes=20, id2word=dictionary)"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[(0,\n",
" '0.007*\"car\" + 0.006*\"increased\" + 0.006*\"closely\" + 0.006*\"groups\" + 0.006*\"center\" + 0.006*\"88\" + 0.006*\"offer\" + 0.005*\"archie\" + 0.005*\"beginning\" + 0.005*\"comets\"'),\n",
" (1,\n",
" '0.005*\"allow\" + 0.005*\"discuss\" + 0.005*\"condition\" + 0.004*\"certain\" + 0.004*\"member\" + 0.004*\"manipulation\" + 0.004*\"little\" + 0.003*\"proposal\" + 0.003*\"heavily\" + 0.003*\"obvious\"'),\n",
" (2,\n",
" '0.002*\"led\" + 0.002*\"mechanism\" + 0.002*\"frank\" + 0.002*\"platform\" + 0.002*\"mormons\" + 0.002*\"concepts\" + 0.002*\"proton\" + 0.002*\"aeronautics\" + 0.002*\"header\" + 0.002*\"foreign\"'),\n",
" (3,\n",
" '0.004*\"objects\" + 0.003*\"activity\" + 0.003*\"manhattan\" + 0.003*\"obtained\" + 0.003*\"eyes\" + 0.003*\"education\" + 0.003*\"netters\" + 0.003*\"complex\" + 0.003*\"europe\" + 0.002*\"missions\"')]"
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
2016-05-26 12:33:27 +00:00
"source": [
"# check the topics\n",
"lda.print_topics(4)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Since there are some problems for translating the corpus from Scikit-Learn to LSI, we are now going to create 'natively' the corpus with Gensim."
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 5,
2019-02-28 14:30:33 +00:00
"metadata": {},
2016-05-26 12:33:27 +00:00
"outputs": [],
"source": [
"# import the gensim.corpora module to generate dictionary\n",
"from gensim import corpora\n",
"\n",
"from nltk import word_tokenize\n",
"from nltk.corpus import stopwords\n",
"from nltk import RegexpTokenizer\n",
"\n",
"import string\n",
"\n",
"def preprocess(words):\n",
" tokenizer = RegexpTokenizer('[A-Z]\\w+')\n",
" tokens = [w.lower() for w in tokenizer.tokenize(words)]\n",
" stoplist = stopwords.words('english')\n",
" tokens_stop = [w for w in tokens if w not in stoplist]\n",
" punctuation = set(string.punctuation)\n",
" tokens_clean = [w for w in tokens_stop if w not in punctuation]\n",
" return tokens_clean\n",
"\n",
"#words = preprocess(newsgroups_train.data)\n",
"#dictionary = corpora.Dictionary(newsgroups_train.data)\n",
"\n",
"texts = [preprocess(document) for document in newsgroups_train.data]\n",
"\n",
"dictionary = corpora.Dictionary(texts)"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 6,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Dictionary(10913 unique tokens: ['cel', 'ds', 'hi', 'nothing', 'prj']...)\n"
]
}
],
2016-05-26 12:33:27 +00:00
"source": [
"# You can save the dictionary\n",
"dictionary.save('newsgroup.dict')\n",
"\n",
"print(dictionary)"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 7,
2019-02-28 14:30:33 +00:00
"metadata": {},
2016-05-26 12:33:27 +00:00
"outputs": [],
"source": [
"# Generate a list of docs, where each doc is a list of words\n",
"\n",
"docs = [preprocess(doc) for doc in newsgroups_train.data]"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 8,
2019-02-28 14:30:33 +00:00
"metadata": {},
2016-05-26 12:33:27 +00:00
"outputs": [],
"source": [
"# import the gensim.corpora module to generate dictionary\n",
"from gensim import corpora\n",
"\n",
"dictionary = corpora.Dictionary(docs)"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 15,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"WARNING:root:random_state not set so using default value\n",
"WARNING:root:failed to load state from newsgroups.dict.state: [Errno 2] No such file or directory: 'newsgroups.dict.state'\n"
]
}
],
2016-05-26 12:33:27 +00:00
"source": [
"# You can optionally save the dictionary \n",
"\n",
"dictionary.save('newsgroups.dict')\n",
2019-04-23 08:39:56 +00:00
"lda = LdaModel.load('newsgroups.dict')"
2016-05-26 12:33:27 +00:00
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 16,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Dictionary(10913 unique tokens: ['cel', 'ds', 'hi', 'nothing', 'prj']...)\n"
]
}
],
2016-05-26 12:33:27 +00:00
"source": [
"# We can print the dictionary, it is a mappying of id and tokens\n",
"\n",
"print(dictionary)"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 17,
2019-02-28 14:30:33 +00:00
"metadata": {},
2016-05-26 12:33:27 +00:00
"outputs": [],
"source": [
"# construct the corpus representing each document as a bag-of-words (bow) vector\n",
"corpus = [dictionary.doc2bow(doc) for doc in docs]"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 18,
2019-02-28 14:30:33 +00:00
"metadata": {},
2016-05-26 12:33:27 +00:00
"outputs": [],
"source": [
"from gensim.models import TfidfModel\n",
"\n",
"# calculate tfidf\n",
"tfidf_model = TfidfModel(corpus)\n",
"corpus_tfidf = tfidf_model[corpus]"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 19,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[(0, 0.24093628445650234), (1, 0.5700978153855775), (2, 0.10438175896914427), (3, 0.1598114653031772), (4, 0.722808853369507), (5, 0.24093628445650234)]\n"
]
}
],
2016-05-26 12:33:27 +00:00
"source": [
"#print tf-idf of first document\n",
"print(corpus_tfidf[0])"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 20,
2019-02-28 14:30:33 +00:00
"metadata": {},
2016-05-26 12:33:27 +00:00
"outputs": [],
"source": [
"from gensim.models.ldamodel import LdaModel\n",
"\n",
"# train the lda model, choosing number of topics equal to 4, it takes a long time\n",
"\n",
"lda_model = LdaModel(corpus_tfidf, num_topics=4, passes=20, id2word=dictionary)"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 21,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[(0,\n",
" '0.011*\"thanks\" + 0.010*\"targa\" + 0.008*\"mary\" + 0.008*\"western\" + 0.007*\"craig\" + 0.007*\"jeff\" + 0.006*\"yayayay\" + 0.006*\"phobos\" + 0.005*\"unfortunately\" + 0.005*\"martian\"'),\n",
" (1,\n",
" '0.007*\"islam\" + 0.006*\"koresh\" + 0.006*\"moon\" + 0.006*\"bible\" + 0.006*\"plane\" + 0.006*\"ns\" + 0.005*\"zoroastrians\" + 0.005*\"joy\" + 0.005*\"lucky\" + 0.005*\"ssrt\"'),\n",
" (2,\n",
" '0.009*\"whatever\" + 0.009*\"baptist\" + 0.007*\"cheers\" + 0.007*\"kent\" + 0.006*\"khomeini\" + 0.006*\"davidian\" + 0.005*\"gerald\" + 0.005*\"bull\" + 0.005*\"sorry\" + 0.005*\"jesus\"'),\n",
" (3,\n",
" '0.005*\"pd\" + 0.004*\"baltimore\" + 0.004*\"also\" + 0.003*\"ipx\" + 0.003*\"dam\" + 0.003*\"feiner\" + 0.003*\"foley\" + 0.003*\"ideally\" + 0.003*\"srgp\" + 0.003*\"thank\"')]"
]
},
"execution_count": 21,
"metadata": {},
"output_type": "execute_result"
}
],
2016-05-26 12:33:27 +00:00
"source": [
"# check the topics\n",
"lda_model.print_topics(4)"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 22,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[(0, 0.09401487), (1, 0.08991001), (2, 0.08514047), (3, 0.7309346)]\n"
]
}
],
2016-05-26 12:33:27 +00:00
"source": [
"# check the lsa vector for the first document\n",
"corpus_lda = lda_model[corpus_tfidf]\n",
"print(corpus_lda[0])"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 24,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[('lord', 1), ('god', 2)]\n"
]
}
],
2016-05-26 12:33:27 +00:00
"source": [
"#predict topics of a new doc\n",
"new_doc = \"God is love and God is the Lord\"\n",
"#transform into BOW space\n",
"bow_vector = dictionary.doc2bow(preprocess(new_doc))\n",
"print([(dictionary[id], count) for id, count in bow_vector])"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 25,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[(0, 0.06678458), (1, 0.8006135), (2, 0.06974816), (3, 0.062853776)]\n"
]
}
],
2016-05-26 12:33:27 +00:00
"source": [
"#transform into LDA space\n",
"lda_vector = lda_model[bow_vector]\n",
"print(lda_vector)"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 26,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"0.007*\"islam\" + 0.006*\"koresh\" + 0.006*\"moon\" + 0.006*\"bible\" + 0.006*\"plane\" + 0.006*\"ns\" + 0.005*\"zoroastrians\" + 0.005*\"joy\" + 0.005*\"lucky\" + 0.005*\"ssrt\"\n"
]
}
],
2016-05-26 12:33:27 +00:00
"source": [
"# print the document's single most prominent LDA topic\n",
"print(lda_model.print_topic(max(lda_vector, key=lambda item: item[1])[0]))"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 27,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[(0, 0.110989906), (1, 0.670005), (2, 0.11422917), (3, 0.10477593)]\n",
"0.007*\"islam\" + 0.006*\"koresh\" + 0.006*\"moon\" + 0.006*\"bible\" + 0.006*\"plane\" + 0.006*\"ns\" + 0.005*\"zoroastrians\" + 0.005*\"joy\" + 0.005*\"lucky\" + 0.005*\"ssrt\"\n"
]
}
],
2016-05-26 12:33:27 +00:00
"source": [
"lda_vector_tfidf = lda_model[tfidf_model[bow_vector]]\n",
"print(lda_vector_tfidf)\n",
"# print the document's single most prominent LDA topic\n",
"print(lda_model.print_topic(max(lda_vector_tfidf, key=lambda item: item[1])[0]))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Latent Semantic Indexing (LSI)"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 28,
2019-02-28 14:30:33 +00:00
"metadata": {},
2016-05-26 12:33:27 +00:00
"outputs": [],
"source": [
"from gensim.models.lsimodel import LsiModel\n",
"\n",
"#It takes a long time\n",
"\n",
"# train the lsi model, choosing number of topics equal to 20\n",
"\n",
"\n",
"lsi_model = LsiModel(corpus_tfidf, num_topics=4, id2word=dictionary)"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 29,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"[(0,\n",
" '0.769*\"god\" + 0.345*\"jesus\" + 0.235*\"bible\" + 0.203*\"christian\" + 0.149*\"christians\" + 0.108*\"christ\" + 0.089*\"well\" + 0.085*\"koresh\" + 0.081*\"kent\" + 0.080*\"christianity\"'),\n",
" (1,\n",
" '-0.863*\"thanks\" + -0.255*\"please\" + -0.160*\"hello\" + -0.153*\"hi\" + 0.123*\"god\" + -0.112*\"sorry\" + -0.088*\"could\" + -0.075*\"windows\" + -0.068*\"jpeg\" + -0.062*\"gif\"'),\n",
" (2,\n",
" '-0.779*\"well\" + 0.229*\"god\" + -0.164*\"yes\" + 0.153*\"thanks\" + -0.135*\"ico\" + -0.135*\"tek\" + -0.132*\"beauchaine\" + -0.132*\"queens\" + -0.132*\"bronx\" + -0.131*\"manhattan\"'),\n",
" (3,\n",
" '0.343*\"well\" + -0.335*\"ico\" + -0.334*\"tek\" + -0.328*\"bronx\" + -0.328*\"beauchaine\" + -0.328*\"queens\" + -0.325*\"manhattan\" + -0.305*\"com\" + -0.303*\"bob\" + -0.073*\"god\"')]"
]
},
"execution_count": 29,
"metadata": {},
"output_type": "execute_result"
}
],
2016-05-26 12:33:27 +00:00
"source": [
"# check the topics\n",
"lsi_model.print_topics(4)"
]
},
{
"cell_type": "code",
2019-04-23 08:39:56 +00:00
"execution_count": 30,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[(0, 0.24093628445650234), (1, 0.5700978153855775), (2, 0.10438175896914427), (3, 0.1598114653031772), (4, 0.722808853369507), (5, 0.24093628445650234)]\n"
]
}
],
2016-05-26 12:33:27 +00:00
"source": [
"# check the lsi vector for the first document\n",
"print(corpus_tfidf[0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# References"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"* [NLTK Book. Natural Language Processing with Python. Steven Bird, Ewan Klein, and Edward Loper. O'Reilly Media, 2009 ](http://www.nltk.org/book_1ed/)\n",
"* [NLTK Essentials, Nitin Hardeniya, Packt Publishing, 2015](http://proquest.safaribooksonline.com/search?q=NLTK%20Essentials)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Licence"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The notebook is freely licensed under under the [Creative Commons Attribution Share-Alike license](https://creativecommons.org/licenses/by/2.0/). \n",
"\n",
2019-03-06 16:46:12 +00:00
"© Carlos A. Iglesias, Universidad Politécnica de Madrid."
2016-05-26 12:33:27 +00:00
]
}
],
"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",
2019-03-06 16:46:12 +00:00
"version": "3.7.1"
},
"latex_envs": {
"LaTeX_envs_menu_present": true,
"autocomplete": true,
"bibliofile": "biblio.bib",
"cite_by": "apalike",
"current_citInitial": 1,
"eqLabelWithNumbers": true,
"eqNumInitial": 1,
"hotkeys": {
"equation": "Ctrl-E",
"itemize": "Ctrl-I"
},
"labels_anchors": false,
"latex_user_defs": false,
"report_style_numbering": false,
"user_envs_cfg": false
2016-05-26 12:33:27 +00:00
}
},
"nbformat": 4,
2019-03-06 16:46:12 +00:00
"nbformat_minor": 1
2016-05-26 12:33:27 +00:00
}
