Research on Collaborative Filtering Book Recommendation Algorithm Based on Clustering and Prediction Fill
As a personalized technology that can improve users satisfaction,recommendation technology is widely used in customer-oriented situations.In this paper,a combined recommendation algorithm is proposed for book rental service in school libraries.The algorithm presets the Mini Batch K-means clustering algorithm on the basis of the common collaborative filtering algorithm,and applies the predicted score data filling technology in the divided reader cluster based on the characteristics of the book items,so as to reduce the sparsity of the reader-book matrix.Therefore,the recommendation efficiency is largely improved.Experiments has revealed that compared with the traditional collaborative filtering algorithm,the recommendation quality of this algorithm has been improved to some extent,and the predictive scoring mechanism can also improve the cold start problem of new books.
Book RecommendationCollaborative FilteringClusteringData FillingMatrix Sparsity
NETL
NSTL
万方数据
