A New Approach to Development of Recommendation Systems with Opinion Mining on Turkish User Reviews

Abstract

People are starting to use Internet more often in their daily lives; they prefer to do more online shopping for their needs. However, e-commerce companies use a variety of recommendation systems to increase product sales. Current recommendation systems are often based on the user’s personal information and online behavior. The most commonly used algorithm in recommendation systems is collaborative filtering. However, despite its wide range of applications, the prediction accuracy of this algorithm is not enough. It can also be questioned whether quantitative data such as product grading or purchasing history reflects users’ true taste. For these reasons, Opinion mining is appeared as a new approach to the development of recommendation systems. In this article we propose a model to analyze Turkish user reviews that can be found a large amount on the internet, based on opinion mining. This model will contribute to the development of recommendation systems.