Title: Mining Heterogeneous Data from Multiple Sources
Dr. Pang-Ning Tan
Department of Computer Science and Engineering
Michigan State University
Date: Friday, November 19, 2010
Time: 11:30 am
Room: 1279 Anthony Hall
Data mining has emerged as an essential part of modern data management. As we continue to be confronted by the ever growing amount of heterogeneous data available, the need for viable data mining techniques that can effectively utilize information from multiple sources is becoming increasingly urgent. For example, satellite-derived remote sensing data, coupled with in-situ observations from weather stations and outputs from global or regional climate simulation models offer great opportunities for scientists to monitor and project future changes in the Earth's climate and ecosystem. In the areas of social and information network analysis, auxiliary information from diverse types of links and nodes can be used to enhance the modeling and analysis of such networks. In this talk, I will discuss some of the challenges involved and present techniques we had developed for mining heterogeneous data in these application domains.
Dr. Pang-Ning Tan is an Associate Professor in the Department of Computer Science and Engineering at MSU. He received his M.S degree in Physics and Ph.D. degree in Computer Science from University of Minnesota. His research interests span a broad range of data mining problems, from pattern discovery (association analysis, anomaly detection, and cluster analysis) to predictive modeling. In addition to addressing fundamental problems in data mining, he is also interested in applying data mining techniques to various application domains including climate and Earth sciences, social and information networks, botnet/webspam detection, and medical informatics. He is the first author of a widely used data mining textbook, which has been translated into Chinese, Korean, and Greek, and is the recipient of the Withrow Distinguished Scholar Award for junior faculty at MSU in 2010.