Fuzzy - TOPSIS based Collaborative Filtering Recommender System in Apache Spark Environment

Author: Sandro Sheklashvili
Keywords: TOPSIS, Recommender System, Apache Spark
Annotation:

Annotation Recommender systems have played a prominent role in online platforms over the last decade. These systems have been incorporated into applications ranging from e-commerce to leisure, successfully enhancing user experience. Moreover, recommender systems are now being applied to a wider diversity of emerging context applications on the Internet including social media and online platforms for communities. In this study, we present a novel collaborative filtering recommender system model. This model differentiates from other recommender system models in that it utilizes fuzzy topsis and uninorm aggregation operator, to compute similarity degrees between users. We demonstrate the application of the proposed model by integrating it in the Movie Online Store platform. The web application is created on Asp.net technology and collaborative filtering method is implemented in Apache Spark environment. The application example illustrates how the proposed model of collaborative filtering recommender system can predict content of interest to users in the platform, based not only on user preferences but also on features of their user profile. The work can be considered as follows: • Learning collaborative filtering and methods used in it. Such as: fuzzy-topsis, uninorm aggregation operator, cosine similarity and etc. • Novel collaborative filtering recommender system implementation. • Demonstrate the work of the algorithm developed by us.