Predicting User Behavior With Tweets Posted In Twitter

Abstract

Twitter is the most well-known microblogging platform around the globe. It has been one of the fastest-growing social network platforms and has a good position in the area of microblogging. More than 500 million registered users post 340 million tweets every day, sharing their thoughts and daily activities. Compared with typical microblogging platforms. Meanwhile, various researches have shown that people tend to express their emotions on Twitter. These tweets usually clearly express user behavior. The most popular emotional analysis task in Twitter is the automatic classification of tweets into sentiment categories such as positive, negative, and neutral. It does not give an idea about user emotion. (e.g., sad, surprise, love). Therefore, our goal is to design an emotion-based algorithm to analyze emotion. To develop an algorithm, we used a dictionary of words for each emotion (lexicon approach). However, we used the support vector machine and naïve Bayes classifiers as machine learning. We have shown that combining these methods provides 82%, 71% accuracy for support vector machine, and naive Bayesian classifiers in the emotional analysis.