gen-text-compr-aggl-clust-sum

A library for topic modeling based on the algorithm: Generative Text Compression with Agglomerative Clustering Summarization (GTCACS)

Installation

Use the package manager pip to install gtcacs.

pip3 install gtcacs

Tested Python version:

python3.8

Tested dependencies:

numpy==1.19.5
scikit-learn==0.24.1
scipy==1.6.1
tensorflow==2.4.1
tqdm==4.58.0

Usage

from sklearn.datasets import fetch_20newsgroups
from gtcacs.topic_modeling import GTCACS

# load dataset
corpus, labels = fetch_20newsgroups(subset='all', return_X_y=True, download_if_missing=False)

# set stop words
eng_stopwords = {'i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', "you're", "you've", "you'll", "you'd", 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', 'she', "she's", 'her', 'hers', 'herself', 'it', "it's", 'its', 'itself', 'they', 'them', 'their', 'theirs', 'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', "that'll", 'these', 'those', 'am', 'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does', 'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while', 'of', 'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during', 'before', 'after', 'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off', 'over', 'under', 'again', 'further', 'then', 'once', 'here', 'there', 'when', 'where', 'why', 'how', 'all', 'any', 'both', 'each', 'few', 'more', 'most', 'other', 'some', 'such', 'no', 'nor', 'not', 'only', 'own', 'same', 'so', 'than', 'too', 'very', 's', 't', 'can', 'will', 'just', 'don', "don't", 'should', "should've", 'now', 'd', 'll', 'm', 'o', 're', 've', 'y', 'ain', 'aren', "aren't", 'couldn', "couldn't", 'didn', "didn't", 'doesn', "doesn't", 'hadn', "hadn't", 'hasn', "hasn't", 'haven', "haven't", 'isn', "isn't", 'ma', 'mightn', "mightn't", 'mustn', "mustn't", 'needn', "needn't", 'shan', "shan't", 'shouldn', "shouldn't", 'wasn', "wasn't", 'weren', "weren't", 'won', "won't", 'wouldn', "wouldn't"}

# instantiate the GTCACS object
gtcacs_obj = GTCACS(
	num_topics=20,                # number of topics
	max_num_words=50,             # maximum number of terms to consider
	max_df=0.95,                  # maximum document frequency
	min_df=15,                    # minimum document frequency
	stopwords=eng_stopwords,      # stopwords set
	ngram_range=(1, 2),           # range for ngram
	max_features=None,            # maximum number of terms to consider (max vocabulary size)
	lowercase=True,               # flag for convert to lowercase
	num_epoches=5,                # number of epochs
	batch_size=128,               # number of documents in a batch
	gen_learning_rate=0.005,      # learning rate for optimize the generative part
	discr_learning_rate=0.005,    # learning rate for optimize the discriminative part
	random_seed_size=100,         # dimension of generator input layer
	generator_hidden_dim=512,     # dimension of generator hidden layer
	document_dim=None,            # dimension of generator output layer and discriminator's input/output layer
	latent_space_dim=64,          # dimension of discriminator latent space
	discriminator_hidden_dim=256  # dimension of discriminator hidden layer
)

# compuation on corpus (dimensional reduction, clustering, summarization)
gtcacs_obj.extract_topics(corpus=corpus)

# get the extracted clusters of words
topics = gtcacs_obj.get_topics_words()
for i, topic in enumerate(topics):
    print(">>> TOPIC", i + 1, topic)

# get the topics distribution scores for each document
corpus_transf = gtcacs_obj.get_topics_distribution_scores()
print(corpus_transf)

License

MIT