基于相似性的商品陈列研究

Commodity display research based on similarity

利用聚类的基本知识,根据不同顾客购买商品的相似性的大小,提出了运用K-means聚类算法。利用相似度代替欧氏距离,对该网络进行聚类分析,划分出相似性大的顾客群体,并根据每个群体中顾客购买每类商品占总商品数的比例进行排序,从而为商品陈列提供依据。

This paper, by using the clustering of basic knowledge, according to different customer to purchase the commodity magnitude of similarity, using similarity instead of euclidean distance, the commodity network clustering analysis, divides the similarity in the types of goods, and according to the customer to purchase each group for each type of goods accounted for the proportion of the number of sorted goods, and thus provides the basis for commodity display.