论文部分内容阅读
[目的/意义]为提高信息服务的针对性和个性化水平,提出一种融合标签权值的用户模糊聚类方法。[方法/过程]根据用户的标注行意义为构造基于各个标签的模糊相似矩阵,然后根据提出的标签权值确定流程来确定不同标签的权值,再在此基础上求得用户标注行为的模糊相似矩阵,最后通过计算确定阈值并根据阈值对用户进行聚类。[结果 /结论]实验结果表明,该聚类算法能够有效地对用户进行聚类,而且聚类准确度要优于传统模糊聚类方法。
[Purpose / Significance] In order to improve the pertinence and individuality of information service, this paper proposes a user fuzzy clustering method based on the weight of tags. [Method / Process] According to the significance of the user’s labeling line, a fuzzy similarity matrix based on each label is constructed, and then the weights of different labels are determined according to the proposed process of determining the weight of the label, and then fuzzy of the labeling behavior of the user is obtained Similarity matrix, and finally calculate the threshold value and cluster the users according to the threshold value. [Result / Conclusion] The experimental results show that this clustering algorithm can effectively cluster users, and the clustering accuracy is better than the traditional fuzzy clustering method.