Applying Data Mining Techniques in Power Engineering: A Case Study

摘要

We pursue the goal of developing a computer aided decision support system that will assist human experts during off-line diagnostic inspections of large rotating electrical machines. The necessary knowledge is supposed to be extracted from a large database where, for the last 35 years, measurement data produced during diagnostic inspections are stored. Since these historical data were not collected for the purpose of data mining, a lot of difficulties need to be elem-inated before the final goal can be achieved. In this paper, we present our experiences made during the data mining process. Particular emphasis is put on the evaluation of the data mining results, which typically must be performed by domain experts. A tool that has been developed for supporting the evaluation of data mining results is presented, and its basic features are applicable to our problem domain.