A Web Recommendation Method Based on User Interest Model

A method of Web recommendation based on user interest model is proposed from the point of mixed recommendation strategy in this paper. For analyzing features of the user interest model, recommendation satisfaction rate under different model parameter C was studied and effectiveness evaluation was done for the proposed method, UAP (by-user-average collaborative filtering predictor) and IAP (by-item-average collaborative filtering predictor) by metric of average absolute deviation (AAD). The optimal C for this interest model is inferred by experimental results which also indicate that the proposed recommendation method performs better than the other two in this research and the problem of data scarcity can be relieved to some extent.