Content Based Sentiment Analysis for Turkish Twitter Feeds Through Vector Space Model

Abstract

Sentiment analysis on social media is the process of determining the sentimental values contained in social media expressions about a specific subject. With the help of sentimental analysis performed by people, institutions or organizations on social media, sentiments in a society related to a certain subject can be determined. These positive, negative, or neutral sentimental values can be used when making decisions about the future that are relevant to the subject. In this study, it is aimed to proportionally determine positive, negative or neutral sentimental values of the tweets written in Turkish on a specific topic on Twitter with the help of Vector Space Model. In order to determine the performance ratio of the work performed, a dataset collected from various social media platforms was created. As a result, classification and sentiment analysis were carried out with 84% accuracy using Vector Space Model.