摘要
为了有效克服传统电子商务用户行为数据聚类存在的不足,提出一种基于密度峰值的电子商务用户行为数据聚类方法.采用小波平稳变换方法对电子商务用户行为数据去噪处理.以局部密度和相对距离的乘积为依据判断电子商务用户行为数据的分布情况,通过正态分布的方式对乘积进行检测,根据密度值获取初始聚类中心和初始簇.利用DBSCAN算法中密度相连的思想,对初始聚类中心聚类处理,获取聚类结果,同时将其划分为对应的类簇内,对相同类簇合并,最终实现电子商务用户行为数据聚类.实验结果表明,所提方法可以有效提升聚类效果和聚类经过,获取满意的聚类结果.
In order to effectively overcome the shortcomings of traditional e-commerce user behavior data clustering,a clustering method of e-commerce user behavior data based on peak density is proposed.Wavelet stationary transform is used to denoise the data of e-commerce user behavior.Judge the distribution of e-commerce user behavior data based on the product of local density and relative distance,detect the product through normal distribution,and obtain the initial cluster center and initial cluster according to the density value.The idea of density connection in DBSCAN algorithm is used to cluster the initial cluster center to obtain the clustering results.At the same time,it is divided into corresponding clusters,and the same clusters are merged to finally achieve the clustering of e-commerce user behavior data.The experimental results show that the proposed method can effectively improve the clustering effect and clustering process,and obtain satisfactory clustering results.
作者
陈金山
Chen Jinshan(Zhangzhou Institute of Science and Engineering)
出处
《哈尔滨师范大学自然科学学报》
CAS
2023年第4期65-69,共5页
Natural Science Journal of Harbin Normal University
基金
漳州理工职业学院横向科研项目“电子商务用户行为大数据研究”(20230101H)
关键词
密度峰值
电子商务
用户行为数据
聚类
Peak density
Electronic Commerce
User behavior data
Clustering
