Emotional analysis towards reasoning about attitudes of customers using Spark — A Novel Machine Learning Approach

Abstract

This paper introduces an information construction methodology to handle eventualities of customer review. The prime attention is on the behavior patterns of the reviewers, building the core for the representation of these eventualities as sequences of emotional and mental actions. Developed foundation facilitates a domain-liberated comparison of the eventualities to manage/analyze behavior patterns of the reviewer. The established model for a special case of interaction (customer reviews) is built and the developed methodology is able to substantiate the sentiment, emotion and reason behind the mental action of the reviewer. The result of this study is that one can assess reviews irrespective of the domain, reasoning about rating of a review and reasoning about the likes/dislikes/emotion associated with the review is also possible. Analysis of emotions of reviewers is done, where the mental conduct is defined in terms of their beliefs and intentions.