New Location Recommendation Technique on Social Network

Abstract

With the availability of current modern technologies, decisions making in an everyday life can be assist in many different ways. Many researches in the past decade has studied about recommendation systems. Recommendation systems can base on different variables with location-based services is one of the more interesting factor to a recommendation system. Recommendations on Location based Network is a service for assisting people to locate locations of their interests. A large number of recorded checked-in histories was gathered to make the prediction according to the desired preferences of each user. Furthermore, determinations have shown a social relationship leading to availability of information will assist in making better recommendations based on the locations. Recently, the recommendation system on location-based domain usually combines either content-based technique and collaborative technique, or collaborative technique and social-based techniques. It is difficult to find the way to combine those three techniques. So there is no research that combine those techniques on location-based recommendation system. This study proposes a new method that combines content-based technique, collaborative technique, and social-based techniques; to produce more efficient result results than location-based RS methods. The evaluation results show that the proposed method provide higher accuracy and coverage than two current location methods by measuring with the Normalized Discounted Cumulative Gain (NDCG) and coverage matrix.