Sentiment Analysis on Textual Data – A comparison of accuracy using different algorithms

Sentiment Analysis on Textual Data - A comparison of accuracy using different algorithms

1Ashish Modi, 2Nambiar Anusree Unnikrishnan

1Assistant Professor, 2Student

1Department of Computer and Information Science,

1Nagindas Khandwala College, Mumbai, India.

1ashishmodi@nkc.ac.in, 2anusreenambiar2003@gmail.com

Abstract— Sentiment Analysis and Opinion Mining is an adverse area of research that deals with the judgement and evaluation criteria in an Emotional communication. Sentiment analysis is generally classified into three different level namely they are document level, sentence level and entity-aspect level. Several research papers are published based on  machine learning techniques such as support vector machine, maximum entropy(maxent) and naïve bayes classifiers and random forest classifiers as the most widely used algorithm in sentiment analysis. Our approach is to take dataset and implement the same using three machine learning models and compare their accuracy.

Index Terms— Sentiment analysis, machine learning, Support Vector machine, Naïve bayes classifier, Sentiment polarity.