A Comparative Study of NLP Topic Modeling Methods and Tools
Shivkumar Goel · International Journal for Research in Applied Science and Engineering Technology · 2019
Analytics today is all about obtaining 'information' from data. Text mining techniques can rapidly gain valuable knowledge and insights from a large amount of unstructured data which is obtained from digital text-based datasets such as web pages, online documents, blogs, articles and emails. Topic modeling is a powerful form of text mining which can be used to extract the data and fetch the information that we are looking for. Topic models represent documents as a 'Bag-of-words' model without taking into consideration the order in which words appear. This paper is meant to study the comparison between four methods which come under the area of Topic Modeling. These methods are Latent Semantic Analysis (LSA), Probabilistic Latent Semantic Analysis (PLSA) and Latent Dirichlet Allocation (LDA). It is also meant to discuss tools available for Topic Modelling-Natural Language Toolkit (NLTK), Gensim and Mallet.