TUTAR, Kadir Ege Üniversitesi - Mühendislik Fakültesi - Bilgisayar Mühendisliği Bölümü, Türkiye , ÜNALIR, Murat Osman Ege Üniversitesi - Mühendislik Fakültesi - Bilgisayar Mühendisliği Bölümü, Turkey , TOKER, Levent Ege Üniversitesi - Mühendislik Fakültesi - Bilgisayar Mühendisliği Bölümü, Türkiye

DEVELOPMENT OF A FRAMEWORK FOR ONTOLOGY BASED SENTIMENT ANALYSIS ON SOCIAL MEDIA

40748

Developing internet technology, trend social media applications and web 2.0 have changed the internet usage habits of internet users. By this means the internet users have started to share their feelings and thoughts on social media from anywhere at anytime. With the increasement of social media usage, valuable feedback data has been increased more and more as well. To this end collection, interpretation and evaluation of this data has come into importance. Sentiment analysis and natural language process methods have been used on text-based data for evaluation and opinion mining to overcome this necessity. In this study, a new ontology-based sentiment analysis method has been developed in order to enhance the accuracy of results that obtained by current sentiment analysis methods. This newly developed method requires to model the domain-specific information on the ontology prior to the analysis procedure. Though this approach, more accurate and more qualified results have been provided to obtain in compared to classic sentiment analysis methods. Another important and innovative feature of this developed infrastructure is being able to do Turkish based sentiment analysis.

194

Sentiment analysis , Ontology , Ontology based , sentiment analysis , Social media feedback management

Pamukkale University Journal Of Engineering Sciences

202