Browsing by Author "Tüfek, Alper"

Now showing 1 - 1 of 1

A novel recommender engine
(Bahçeşehir Üniversitesi Fen Bilimleri Enstitüsü, 2014-04) Tüfek, Alper; Mimaroğlu, Selim Necdet
Recommender systems are software tools and techniques that help users to find products/items which are of interest, from large catalogs. Available options extremely differ both in number and attributes depending on the domain, that is, the type of object/item needed to be selected. Recommender systems can be classified broadly into three categories: content-based, collaborative filtering based and hybrid systems. Content-based systems generate recommendations based on descriptions or content of items. The user will be recommended items similar to the ones the user preferred in the past. The biggest limitation of content-based techniques is that extracting features associated with items to be recommended is usually a costly process. The content must either be in a form that can be parsed automatically (e.g., text) or the features should be assigned to items manually. Collaborative filtering is the most popular technique for recommender systems. Recommender systems of this group simulate taking recommendations from friends with similar tastes. In this thesis, a novel recommender system based on collaborative filtering is designed which can be easily applied to many different domains. The main advantage of the new system is its ability to use both implicit and explicit information which considerably increases recommendation coverage. Also an asymmetric approach is proposed for similarity calculations during nearest neighbor selection procedure. Another objective that is aimed to observe is to be better at differentiating especially liked items from disliked ones. In this respect, a penalization scheme is incorporated to lower down the scores for items with low ratings whereas highlighting items with high ratings.