A Survey on Sentiment Analysis and Topic Modeling

Abstract

Sentiment Reason Mining is an emerging research area in this era of social media. Sentiment Reason Mining aims to resolve two problems: first is finding the reason of a sentiment, and second is interpreting sentiment variations. Time and Event where sentiment is being changed is also an important factor. Aspect-Based methods, Supervised Learning, Topic Modeling, and Data Visualization etc. can be used for finding the reason of a sentiment. VADER Sentiment Classifier can be used for sentiment of tweets. LDA is topic Modeling algorithm. In this research paper we have reviewed some the research work performed for this purpose. We have reviewed various research work which have used social media content as dataset. TF/IDF feature extraction is used in most of the work. Sentiment Detection tools VADER and Text Blob are also discussed in our work.