Recommendation ranking based on resource allocation dynamics on bipartite graph
Hongyan Zhang · Jisuanji gongcheng yu sheji · 2010
In order to meet the requirements of precision and generalization power of internet information recommendation system,the use of resource-allocation dynamics on user-object bipartite graph to learn the recommendation power of each user and object in the unsupervised manner is proposed.The proposed scheme mines the history data of user and item correlations and automatically extracts subjective knowledge on user interests.Experiments demonstrate that,by restraining the power of user or object with large degree,proposed algorithm achieves effectiveness and efficiency,and improves the performance by a margin of 20%.