内容简介：
As websites continue to grow in size and complexity, web usage mining systems face three significant challenges – computational efficiency, accuracy and scalability. This talk will first present our research in the development of relational fuzzy subtractive clustering (RFSC), an efficient and scalable technique for mining from web log data.  This will be followed by describing an incremental version, a generalization technique, and methods for dependable evaluation of clustering results and application tasks. We will conclude with performance results of our experiments on large real weblog data which demonstrate improved performance in terms of quality and computation time.

报告人介绍：
Dr. Sudhir Mudur, is professor and chair of the computer science and software engineering department at Concordia University in Montreal, Canada. His primary area of research is computational techniques for 3D graphics and for analysis of very large data sets. In his research career of over 33 years, he has published extensively in major international journals and conferences.