用户和项目联合度对二分网络个性化推荐的影响

首先采用物质流动算法进行二部图相似系数投影，然后利用随机游走模型得到协同过滤结果。在计算相似系数时，采用了考虑用户和项目联合度分布特征的改进算法。通过数据模拟可知，在最优情况下推荐项目准确率提高了18. 19%，推荐项目多样性提高了21. 90%。对用户和项目联合度的分布进行了统计分析，结果表明，在最优情况下，其符合指数为--2. 33的指数分布。

Effort of User-Item Degree Correlations to Bipartite Network Personalized Recommendations

In this paper first bipartite graph was project based on mass diffusion, then random walk method was used to get collaborative filtering results. Degree correlation between users and objects was embedded into the similarity index to improve the algorithm The numerical simulation shows that the algorithmic accuracy of the presented algorithm is improved by 18. 19% in the optimal case and the diversity is improved by 21. 90%. The statistical analysis on the prodtrct distribution of the user and object degrees indicates that, in the optimal case, the distribution obeys the power-law and the exponential is equal to--2. 33.