A Knowledge Management Platform for Optimization-based Data Mining

Abstract

Multiple criteria linear programming (MCLP) approach to data mining has been used in many fields. But users need to understand well with math and technology in its working process. This prevents it from wide applications. Studied on standards of data mining process and its advantages to project operation with the analysis on the characters of MCLP method and its process, we found the current data mining process model can not support MCLP in detail. So a knowledge management platform was presented for standardization the MCLP process by referring to CRISP-DM and the researches on data mining process models. The platform collects the experts? experience in daily works of data mining and then accumulates knowledge for standardization. Its application in a web company shows that it makes easier for different types of users to work in optimization-based data mining process.