A Service Recommendation Algorithm Based on Modeling of Dynamic and Diverse Demands

摘要: This article contends that current service recommendation algorithms are still unable to meet the dynamic and diverse demands of users, so a algorithm considering is proposed. The latent Dirichlet allocation model machine learning field adopted extract user implicit demand factors, then bipartite graph modeling random-walk used extend factors predict short-term changes diversity demand. At last, list generated based on these factors. Experimental results real-world data set regarding composition show proposed can represent demands, performance better than other in terms accuracy, novelty, diversity.