A Feature-based Personalized Recommendation System In E-commerce: Design And Implementation

With the fast development of Internet, E-commerce has become more and more popular all over the world. Business hence could overcome spatial and temporal barriers and are now capable of serving customers electronically and intelligently. However, the exponentially increasing amount of data and information along with the rapid expansion of business web sites and information systems make business hard to manage. On the other hand, it is also difficult for customers to find the products they want. For these reasons, the personalized recommendation system arises at the right moment, which provides customers one-to-one service based on their past behavior and reference from other users with similar preferences. Many companies nowadays are using this system to retain existing customers and attract new ones.Despite significant progress in personalized recommendation system research, there are several problems that limit its application to E-commerce. Taking the main ones into consideration, this thesis carries out researches and explores some key technologies of recommendation system. The main research results of this thesis are as follows:1) Feature-based Association rule Model (FARM). The expansion of E-commerce web sites and information systems makes items hard to calculate and slows down the systems. To address this issue, we propose a feature-based association rule model, which is based on product features. Through customers' transaction records, their preferences toward specific product features can be learned, and the problems generated by traditional approaches can be solved thereby.2) Feature-based Collaborative Filtering Model (FCFM). The traditional method brings two problems. Firstly, the system performs mass product comparisons prior to finding similar communities for target customers. Secondly, products not yet purchased and rated could not be recommended to customers. To address these issues, we propose feature-based collaborative filtering model. This method first clusters items by their features. Based on the similarity between target items...