Recommender System with Configuration in Electronic Commerce

초록

There are recently a number of techniques to give customers facilities such as reverse auction, common purchase, and recommender systems for using customers’ profiles. However, customers face many problems that can arise when they buy some products requiring technical knowledge. In this paper, we have applied a configuration technique to solve these problems. We have focused on the recommendation of the proper product by analyzing precisely customers’ requirements and on the improvement of the system’s performance in configuration. For recommending proper products to customers, we have designed the systems for a product model that processes an interactive query. Rule-based systems are applied for the interactive query. For the improvement of the system’s performance, we have constructed a partial solution product based on various requirements of the customers. We use algorithms of CSP for this. By simultaneously updating the product information with current product stock of suppliers, we allow the solution product to be configured correctly and to be bought in time. We have implemented the configuration system for assembling the computer by applying on the web.