异质网络资源多维度推荐模式研究——以豆瓣网为例

Multi Dimension Recommendation Model of Heterogeneous Network Resources-A Case Study of Douban

[目的 /意义]针对某些包含多级用户和多级资源的异质网络,在总结其各种异质模态的基础上提出一种多维度推荐框架MDRM,向目标用户进行好友和资源的推荐。[方法/过程]通过对用户、资源划分等级,在分析各种网络模态的基础上建立其异质关系;利用情感倾向分析得到二级用户——二级资源评分矩阵,借助协同过滤算法,实现同级用户和二级资源的推荐;基于异质关系,实现一级用户和一级资源的推荐,最终实现多维度推荐。[结果 /结论]在以豆瓣网数据作为数据集的实验中取得了较好的效果,说明MDRM模型适合某些异质网络资源的推荐。

[ Purpose/significance ] For heterogeneous networks containing multi-level users and multi-level re- sources, a variety of heterogeneous modes are summarized. On this basis, this paper proposes a multi-dimensional recom- mendation framework to recommend friends and resources to the target users. [ Method/process] Firstly, we establish the heterogeneous relation among users and resources by grouping and analyzing all kinds of network models of them. Second- ly, we recommend users of same level and secondary resources by using collaborative filtering algorithm, which is based on the secondary resource score matrix by sentiment analysis. Thirdly, we recommend primary users and resources based on heterogeneous relation, [ Result/conclusion ] The experiment results on douban data show that the proposed recommenda- tion frame is suitable for the recommendation of some heterogeneous network resources.