Topic Modeling Algorithms (TMA) widely used in the field of Aspect Based Review Analysis and demonstrated good performance. The prime intuition behind topic models is that each document is a collection of some topics. A topic is a collection of words that represent the topic as a whole. TMA extract the ‘latent’ semantic topics and themes in a collection of documents. Latent Dirichlet Allocation (LDA) is the most popular TMA used in various text mining applications. In this study, we give deep insight on LDA and its combination with other approaches in Opinion Mining or Sentiment Analysis domain. The purpose of this paper is to provide an ample analysis of various extensions and combinations of LDA for optimizing the complete process of review miming.
展开▼
