Modeling and analysis of a dynamic user behavior and social trust relationship for eCRM

Abstract

In the last decade since the first research about customer behavior in online environment was originally published, there has been much research efforts both in industry and academia on modeling dynamic user behavior and analyzing groups or personal behavior patterns in B2C environment. From Customer Relationship Management (CRM) perspective, it is critical to understand and model the change of customer’s behavior patterns (i.e., purchasing patterns, browsing patterns or reaction patterns toward a promotion) based on which a company will be able to promote desirable trends and prevent undesirable trends. In recent years, the advent of Web 2.0 environments and related technologies have brought the new digital age such as P2P environment (e.g., Web-based social networks) since the advent of World Wide Web which also created a new society and a new industry. In Web-based social networks, a vast majority of current online users are encouraged to share their personal profiles, personal interests, knowledge or valuable user-created contents with anonymous users. However, online users in such Web-based social networks face a challenge to find a large number of trustworthy users from anonymous users. Therefore, for the success of online communities, it is vital to provide a reliable computational trust model which assesses a degree of trust on anonymous users in the eye of an individual user. The eventual goal of a trust model for P2P environment is the same with a dynamic user behavior model for B2C environment, in terms of Customer Relationship Management concerns attracting and keeping valuable customers. However, the unique nature of a B2C online market such as Amazon.com and a P2P based online community makes great differences. Thus, there is need to model and predict consumer (online user) behavior with a different approach right for each market with available data. In this thesis, we suggest dynamic user behavior models for a B2C traditional e-commerce mar...