Inferring values of recommendation links: Analysis of co-purchase network based on ERGM and product involvement

Abstract

Co-purchase based recommendation systems are widely implemented and testified its efficient performance in various kinds of industries. Previous work on measuring the values of co-purchase recommendation links lack of consideration about exogenous attributes of products or their sample are subjectively set the degree of product involvement. The purpose of this paper is to infer values of recommendation links of co-purchase network based on the implicit exogenous product involvement. The statistics network inference model, Exponential random graph models (ERGM) is exploited to analyze the effects of product involvement, recommendation links, and in-coming effect. Our findings show how users co-purchase products in the perspective of both aggregate-and individual-product levels, providing theoretical fundamentals for the design of recommendation systems and the cross-selling marketing strategies.