Clustering Problems in Business Analytics
Dr. Jianying Hu
Business Analytics and Mathematical Science
IBM T. J. Watson Research Center
Friday, March 6
10:00 AM - 11:00 AM
3105 Engineering Building
Host: Anil K. Jain
Business analytics is defined as the gathering and interpretation of data to assist business decision making and business process optimization. The increasing amount of data being collected on a daily basis, from business operations to performance metrics to employee work records, presents a great opportunity for advanced research in pattern recognition, data mining and statistical modeling to create significant business value. At the same time, they also present unique challenges in these areas. This talk will focus on the clustering problems encountered in business analytics. I will first give a brief overview of the business applications we have encountered where clustering analysis provides critical value. I will then present three recent projects on problems inspired by business applications: K-means clustering of proportional data using L1 distance, categorization using semi-supervised clustering, and co-clustering with dual supervision.
Dr. Hu studied electrical engineering at Tsinghua University in Beijing, China, and came to the United States in 1988 for graduate study at SUNY Stony Brook. She received her Master's degree in 1991 and Ph.D. in 1993, both in Computer Science. Her Ph.D. advisor was Prof. Theo Pavlidis. During her graduate study she worked on conic splines, road tracking in aerial images, and shape based curvilinear object indexing and retrieval. She was at Bell Labs from 1993 to 2000, and Avaya Labs from 2000 to 2003. Dr. Hu joined IBM T.J. Watson Research Center in June 2003, initially working on pen based technologies. Her current work focuses on statistical analysis for business analytics. She is a senior member of IEEE.