題 目：Class Restricted Clustering and Micro-Perturbation for Data Privacy
主講人：Xiaobai Li教授（Manning School of Business at the University of Massachusetts Lowell, USA）
時 間：2012.6.25（周一）下午2:30
地 點：主樓418會議室
主講人簡介：
Dr. Xiaobai Li is a Professor of Management Information Systems in the Department of Operations and Information Systems, Manning School of Business at the University of Massachusetts Lowell, USA. He received his Ph.D. in management science from the University of South Carolina in 1999. Dr. Li’s research focuses on data mining, information privacy, and information economics. He has received funding for his research from National Institutes of Health (NIH) and National Science Foundation (NSF), USA. His work has appeared or is forthcoming in Information Systems Research, Management Science, Operations Research, IEEE Transactions on Knowledge and Data Engineering, IEEE Transactions on Systems, Man, and Cybernetics, IEEE Transactions on Automatic Control, Communications of the ACM, Decision Support Systems, INFORMS Journal on Computing, European Journal of Operational Research, among others.
內容簡介：
Data mining and sharing technologies have enabled organizations to extract useful knowledge from data in order to better understand and serve their customers, and thus gain competitive advantages. While successful applications of data mining are encouraging, there are growing concerns about invasions to privacy of personal information by information technology in general, and by data mining/sharing in particular. A variety of approaches have been proposed to resolve the conflict between data mining/sharing and privacy protection. This seminar provides an overview of the current state-of-the-art in this cutting-edge research area, and introduces an innovative approach for data privacy. Using a minimum spanning tree technique, the proposed approach clusters data such that the data points within a group are similar in the non-confidential attribute values whereas the confidential attribute values within a group are well distributed. A novel cluster-level micro-perturbation method for masking data is also proposed, which is shown to preserve the statistical properties of the data.This research area is cross-disciplinary in nature. It has attracted researchers from broad areas of background, including information systems, computer science, statistics, operations research, and economics. We will introduce various approaches from different perspectives, and explain how these different approaches interact and complement each other.
