Research on product novelty recommendation based on user demands

The recommender systems play an important role in alleviating information overload, extracting interesting commodities for users and improving user personalized experience. Today, with the increasing abundant of products and the personalized needs of users, the tradition recommendation systems fail to enhance user satisfaction because they overemphasis the accuracy of results. The recommendation researches have increasingly focused towards introducing novelty in user recommendation lists. Existing methods aim to find the right balance between the similarity and novelty of the recommended items. However, they ignore the different user demands for the accuracy and novelty. Therefore, this paper further analyzes user characteristics according to product types and quantity selected by users, constructs users' demands using the information entropy theory and proposes an adaptive random walk model based on user demands. The experimental results show that the proposed model can adaptively meet users' needs for novelty while ensuring accuracy and enrich the models of the novelty recommendation.