报告题目：Single Model versus Matching Model
报告人: Dr. Kai Ming Ting
Gippsland School of Computing and Information Technology, Monash University, Australia
报告地点: 蒙民伟楼404
报告时间: 2005年9月19日 14:30
摘要:
A tacit assumption in classifier induction is that the class distribution of the training set must match the class distribution of the test set. A direct implementation is to retrain a model using a data set with matching class distribution every time the operating condition changes (i.e., the matching model). The alternative is to modify the decision rule of a previously trained model to the new operating condition. The latter is the single model approach commonly used and recommended by many researchers.
In this talk, I report, with empirical support using decision trees, that learning using the matching class distribution is desirable. I also make explicit the differences and limitations of the two methods for the single model approach: rescaling and thresholding.
报告人介绍:
Kai Ming Ting博士, 1986年于University of Technology Malaysia电子工程系获得学士学位，1992年于University of Malaya计算机系获得硕士学位，1996年于University of Sydney计算机系获得博士学位。2002年至今担任澳大利亚Monash大学Gippsland学院计算与信息技术系高级讲师、副系主任。曾担任ICML、ECML、KDD等多个重要国际学术会议的程序委员会成员, 以及JAIR, MLJ, JMLR, TKDE, TPAMI, TIS等10余家重要国际刊物的审稿专家。主要研究领域为机器学习、数据挖掘。
|
