文档介绍：effectively relieved the sparsity of rating data and improved the accuracy of the neighbors selected. For the third point, the influence of both of user context information and user rating information are taken into consideration, users’ comprehensive similarity is got by a certain bined with users’ subjective rating similarity and objective context similarity. This makes the system can mend for new users who have never evaluated any item.
This article used both metrics of mean absolute error(MAE) and mean absolute percentage error(MAPE) to verify the algorithm on MovieLens dataset. The experimental results indicated that, compared with traditional collaborative filtering mendation algorithm and an improved collaborative filtering mendation, the mendation accuracy of the advanced approach was improved to some extent. Lastly, the article illustrated the practical application of the modified method through an actual project.
Keywords: User Context, Fuzzy Clustering, Slope One, Collaborative Filter, mendation
目 录
中文摘要..........................................................................................................................................I
英文摘要........................................................................................................................................ II
1 绪 论......................................................................................................................................... 1
研究背景................................................................................................................................... 1
国内外研究现状....................................................................................................................... 2
论文研究的目的和内容........................................................................................................... 3
论文研究的目的................................................................................................................ 3