Sentiment Analysis on Turkish Social Media Shares through Lexicon Based Approach

Social media platforms provide an environment that allows users to see the shares made up to that time on a particular subject or situation. Reading and analysing millions of comments made on a given subject or situation is a costly process that takes considerable amount of time. For this reason, the development of applications that automatically perform such analyses has become a necessity nowadays when the use of social media is increasing rapidly. In this study, messages written in Turkish that had been shared on Twitter, which is one of the most used social media platforms, were analysed with the help of lexicon-based method which is one of the approaches used in sentiment analysis after being passed through various pre-processing stages. As a result of this sentiment analysis, according to the sentimental densities they carry, they have been classified in three categories namely positive, negative, or neutral. As a conclusion of the studies performed, the classification and sentiment analysis process was performed with a success rate of approximately 80%.