A machine learning approach to extract opinions from social media content

Abstract

The Opinion Mining (OM) from mobile based social media content (SMC) is more challenging compared to topic-based mining, and it cannot be performed based on just examining the presence of single words in the text containing opinion expressions. Moreover, the existing systems of opinion classification find that a large number of features that are not feasible for the mobile e nvironment. The existing methods of OM in this mobile environment do not consider the semantic orientation of the SMC in the r eview. The proposed machine learning approach extends the feature-based classification approach to identify the orientation of the phrase on taking context into account to improve the accuracy.