Research Area:  Machine Learning

Abstract:

Sentiment analysis aims to reveal semantic knowledge of written texts where users share their feelings and thoughts on sharing platforms such as personal blogs and social networks. The data shared by users on social networks consist of short texts. A text classification problem has arisen for data reaching large dimensions shared on social networks. Although language libraries have been developed in other languages to solve the problem of sentiment analysis, studies for the Turkish language are limited. In this study, two categories of emotion analysis were studied with data obtained from many social networks. Also, there is a feeling class on a topic on Twitter, and text is summarized in the class. Sentiment analysis considered a classification problem. To increase the success rate, the study was carried out by focusing on the words with semantic context word embedding methods. LSA is used for text summarization. In this study, where both emotion analysis and text summation are carried out, the main goal is to analyze the sentiments and thoughts about a subject and present brief information to the user. The primary model was created with data collected from many social networks. Analysis and summary of the text were made with data from a hashtag on Twitter. The methods used in the analysis of emotions were compared to the methods of word embedding and a success rate of 93% was obtained.

Keywords:  

Author(s) Name:  Emre Doǧan; Buket Kaya

Publisher name:  IEEE

DOI:  10.1109/IDAP.2019.8875879

Volume Information:  

Paper Link:   https://ieeexplore.ieee.org/abstract/document/8875879