Computational Exploration of Theme-based Blog Data Using Topic Modeling, NERC and Sentiment Classifier Combine

Abstract

This paper presents findings of our exploratory research work on a novel combine of Topic Modeling, Named Entity Recognition and Sentiment Classification for sociological analysis of blog data. We have collected more than 500 blog posts on the broader theme of ‘Discrimination, Abuse and Crime against Women’. We employed topic discovery to identify top keywords and key themes and implemented the 7-entity model Named Entity Recognition process to identify the key persons, organizations and locations discussed in the blog posts. Thereafter we performed sentiment classification of the entire blog data into positive and negative categories using SentiWordNet. The results obtained are very interesting and validate the usefulness of our approach for computational analysis of social media data. The key contribution of the paper is to propose a novel Text Analytics combine and demonstrate its applicability for computational exploration of the social media data for sociological analysis purposes.