A comparative study of classifier algorithms for Twitter’s sentiment based spam detection

Abstract

In today’s time Social media is a vital part of everybody’s life. People are expressing their feeling and emotions on social media platforms. Emotions are associated with the daily life experiences of everyone. Twitter is one of the mostly used social media platform. From analysis on social media, we can predict the mental status of users. In this paper a comparative study of user’s posts on Social media has been done. The approach is based on relevant keywords. Sentiment analysis and classification of emotions is done using Bayes Network Classifier, Naive Bayes Classifier, Logistic Regression, Simple Logistic, SMO, J48 pruned tree, and Random Forest. Finally, accuracy of classification is evaluated by different classifiers.