Recommendation of e-commerce sites by matching category-based buyer query and product e-catalogs

Abstract

In this paper, an e-commerce site recommendation system that integrates multiple e-commerce sites is proposed. This system provides the users with a unified portal through which the users can search individual suppliers’ product categories efficiently. The core part of the system is an intelligent product meta-search engine that has the following functions: (1) it provides category-based query, with which a buyer can describe his or her product search intention using superclass/subclass relationship, (2) by using WordNet, the buyer's query is semantically extended in order to increase product search accuracy, and (3) the meta-search engine decides a recommended priority of the suppliers by matching the buyer's query with the suppliers’ product categories and computing a semantic relevancy measure. Experiments show that the performance of the meta-search is better than those of general keyword-based search and category-based search.