A Hybrid Model for Online Merchandise Recommendation Based on Ordination and Cluster Analysis

摘要

With the continuous development of modern information technology, online shopping is becoming more and more popular. With more and more online products, how to rank and recommend online products is particularly important. This paper proposed a hybrid model combining unconstrained ordination analysis and cluster analysis. Ordination analysis is used to explain the relationship between online merchandise and its indexes; then cluster analysis is implemented to classify the results of the ranking analysis. Therefore, the buyer can directly understand the situation of the product in terms of the product index and the store index from the bi-plots. The proposed model solves the neglect of the link between commodities and their indexes in traditional rankings. Buyers can purchase goods accurately according to their needs.