Segmentation based Twitter Opinion Mining using Ensemble Learning

Abstract

In recent years, social media has become the prime place for advertisements, activities, campaigns, protests etc. It provides a platform for the people to express their views and beliefs to the masses. The user beliefs, practices and interests are of great importance to organizations and provides insight into the minds of users. Data Mining is one such tool that enables these organizations to extract relevant information from user data, which can be analyzed to create a knowledge set and determine user opinion that allows companies to create products tailored to the user. Data Mining of Twitter and other social platforms is of a great importance because, its large user base is a goldmine of public opinions and views which if analyzed properly, can potentially be used to predict campaign results and product assessments and likeability. This project proposes a classification scheme that aims to perform Segmentation based Twitter Opinion Mining using Ensemble Learning. The proposed scheme is able to detect and filter out bots and uses text segmentation for effective text classification and part of speech tagging.Keywords - machine learning, supervised learning, text analysis, sentiment analysis, natural language processing.