A Survey of Sentiment Analysis of Internet Textual Data and Application to Pakistani YouTube User Comments

Abstract

With the advent of social media, e-commerce and video sharing websites, a large amount of textual data depicting people’s opinions is being generated. This text has a great impact on other web users, product sellers and content creators. Text mining and emotion detection using text analysis has prospered a lot in recent years. Sentiment Analysis is the method of mining emotions, reviews and opinions of web users. In this paper, a survey of text Sentiment Analysis techniques is presented. Sentiment Analysis is applied to comments extracted from Pakistani news videos. Different machine learning algorithms previously surveyed in this paper have been applied to this dataset. Our results have shown that transfer learning can be effectively used for Sentiment Analysis.