- Avrupa Bilim ve Teknoloji Dergisi
- Ejosat Special Issue: (HORA) Special Issue
- Turkish Sentiment Analysis System via Ensemble Learning
Turkish Sentiment Analysis System via Ensemble Learning
Authors : Saed Alqaraleh
Pages : 122-129
Doi:10.31590/ejosat.779181
View : 11 | Download : 8
Publication Date : 2020-08-15
Article Type : Research
Abstract :Nowadays, sentiment analysis (SA) also known as opinion mining (OM) is widely used and has an impressive effect in many fields, such as marketing, politics, and even, company’s products are now adjusted based on users’ opinions. In this paper, a new efficient sentiment analysis system that supports the Turkish language has been introduced. In addition, as Turkish is an agglutinative language, which requires special processing, an efficient preprocessing model was also implemented as a part of the developed system. Several experiments using the challenging and benchmark "The Turkish movie reviews” dataset have been conducted, and it is obvious that the constructed approach can efficiently support the Turkish language and can achieve a quite good performance.Keywords : Sentiment Analysis, Text Classification, Opinion Mining, Turkish Language, Ensemble learning, Natural Language Processing