To Provide Personalized Services in Mobile Commerce using Distributed Bayesian Networks

Project: Research

View graph of relations


This project will propose a new distributed Bayesian network-based personalization model and implement the model in an existing m-commerce application platform to prove its efficiency and effectiveness. To provide personalized services, the users/services need to be identified in an accurate way, which requires an effective representation of the information relationships and the associated inferring abilities. Data mining techniques such as Bayesian networks have already been proven effective tools to support personalization in m-commerce. However, m-commerce data, including user profiles, user preferences, context, and content information, are by nature very distributed, privacy sensitive, and heterogeneous. Hence, mining all data at a centralized place is usually not allowable in m-commerce due to both managerial and technological problems/difficulties. It would be more feasible and effective if data mining could be performed in a distributed way. Moreover, using a distributed Bayesian network offers a promising way to tackle other special features of m-commerce data, such as the data’s inaccuracy, incompleteness, and ambiguity. To enhance the effectiveness of data mining and personalization in m-commerce, the distributed Bayesian network and its implementation are of great importance. However, little research has been done so far for this purpose.


Project number9041287
Grant typeGRF
Effective start/end date1/08/0728/10/10