Application of Recommender Systems on Web Portals to Increase User's Visit Statistics

Abstract

In today's competitive world, the rapid advancement of technology, Internet and electronic commerce raise demand for existence of a mechanism which can predict user’s requirements and requests. These mechanisms can lead us to outsource our competitors. The main issue is that we encounter large amount of information in web portals which are mostly heterogeneous and unrelated, therefore with no proper strategies of categorizing data and information preparation, users get involved in confusion accessing correct content. The most important challenges are reaching most relevant information in order to provide users. As a matter of fact, this problem could get solved by using domain of recommender systems which can help us finding and selecting related information according to user needs. Although, recommender systems help people dealing with massive data, these systems are less employed on Web portals. Certainly, the application of these types of systems on web portals will bring decent improvements to users. The study, uses MovieLens20m dataset that includes ratings and user labels for movies. User ratings and movie rating relationships are used in order to make an appropriate recommendation for users. labels which users considered for the movies are also employed in extraction phase. Finally, a combination of these two categories is offered as a recommendation to user.

Keywords: Recommender Systems, Web Portals, User Visits, MovieLens

DOI: 10.7176/JIEA/9-3-03

Publication date:May 31^st 2019