Implicit behavior addition for improving recommendations

Recommender system is used in e-commerce websites. An e-commerce website should capture the customer´s preferences at best to recommend her well. The core area around which a recommender system is build is user profiling. Researchers and industry professionals are working in the direction of user profiling to improve the recommendations. This paper is an attempt to closely study the impact of adding implicit behavior of a user in recommendation results generation. It highlights the conceptual overview of some implicit behavior based attributes which are useful in producing quality recommendations through reordering. In Context of Recommender system Top-N occupancy is of high importance. Products getting slot in Top-N occupancy are highly likely to be converted to sales and gain users´ trust. Considering this fact, we are proposing an implicit behavior based solution which mainly focuses on Top-N occupancy. It also presents an approach through which similarity in implicit behavior based attribute is calculated.