NMF topic modeling

In this recipe, we will use another unsupervised topic modeling technique, NMF. We will also explore another evaluation technique, topic model coherence. NMF topic modeling is very fast and memory efficient and works best with sparse corpora.

Getting ready

We will continue using the gensim package in this recipe.

How to do it…

We will create an NMF topic model and evaluate it using the coherence measure, which measures human topic interpretability. Many of the functions used for NMF models are the same as for LDA models in the gensim package. The steps for this recipe are as follows:

1. Perform the necessary imports:

   import re
   import pandas as pd
   from gensim.models.nmf import Nmf
   from gensim.models import CoherenceModel
   import gensim.corpora as corpora
   from gensim.utils import simple_preprocess
   import matplotlib.pyplot as plt
   from pprint import pprint
   from Chapter06.lda_topic_sklearn import stopwords, bbc_dataset, new_example
   from Chapter06...