Modeling with Gensim

The Gensim’s LSI module implements fast, truncated SVD. The module is actually engineered to process very large corpora through distributed computing. It can also take a stream of corpora as input. Let’s use the LsiModel() function to build our first model. We will build it with the corpus in BoW and TF-IDF.

BoW

Let’s start with the model with the corpus in BoW:

import gensimlsi_model = gensim.models.lsimodel.LsiModel(
      corpus=bow_corpus, 
      id2word=gensim_dictionary,    
      num_topics=20)

Let’s explain the previous parameters:

• corpus: This is the corpus of our data. It is a required input parameter. Here we use bow_corpus.
• id2word: This is a list of the IDs of the words. We assign our dictionary as gensim_dictionary. If we do not assign a dictionary, Gensim will show a warning message but...