摘要
针对协同过滤算法忽视供应商偏好、存在稀疏矩阵导致准确率低的现象,提出一种改进的协同过滤算法。利用改进的相似度计算方法填充评分矩阵,计算目标用户的评分,将目标用户评分作为G-S算法的输入项,得到消费者、供应商的匹配结果。仿真结果表明,算法具有较高的满意度和准确率。
The absence of supplies' interest and the execution efficiency of recommendation technology based on collaborative filtering algorithm is relatively low with the data sparsity,so a modified collaborative filtering algorithm is proposed.Firstly,it prefills the rating matrix by the approved algorithm and makes sure of the users' rates.Then the improved G-S algorithm for consumers and sellers provides appropriate match on both sides according to the recommended goods based on the collaborative filtering algorithm.The experimental results show that the algorithm has high execution and satisfaction.
出处
《桂林电子科技大学学报》
2015年第5期395-400,共6页
Journal of Guilin University of Electronic Technology
基金
国家自然科学基金(61262074)
关键词
协同过滤算法
满意值
PARETO最优
信息熵
collaborative filtering algorithm
satisfaction
Pareto principle
information entropy