New dimensions of temporal serendipity and temporal novelty inrecommender system
Chhavi Rana · Advances in Applied Science Research · 2013
Recommender system focuses on techniques that could predict user interest and give assistance while th e user interacts with the Web in finding relevant informat ion. It attempt to make sense of the data generate d by his past interaction and predict in future choices. The focu s of research in the area of recommender system has been on accuracy in the past decade, but the trend is chang ing with an increasing interest in this area of res earch. This paper is an attempt to provide an overview of the sof the art in new dimension of recommender sys tem research. Novelty and serendipity refers to the search of fin ding something new by a user while browsing world w ide web. Traditional recommender system algorithm focuses on accuracy that tries to compare accuracy with past data which limits the scope of novelty and serendipity tgreat extent. Novelty pertains to giving someth ing new which the user have not accesses before but similar in ta ste while serendipity is a chance discovery that co uld be really beneficial for a user at certain times. This paper will present an outlook on the existing research ca rried out in this area, their specialized focus with respect to an ap plicative objectives and the need for a more compre hensive new entrant in this sphere in the light of the current scenario. The paper will also present a novel metho dology based on temporal parameters to include the novelty and sere ndipity in recommender system. In the end, the pape r will be concluded by listing some challenges and future tre nds in this research area.