SVM-Based Multiclass Cost-sensitive Classification with Reject Option for Fault Diagnosis of Steam Turbine Generator

Abstract

The steam turbine generator faults not only damage the generator itself, but also cause outages and loss of profits, for this reason, many researchers work on the fault diagnosis. But misdiagnosing may also lead to serious losses. In order to improve the diagnosis reliability and reduce the loss caused by misdiagnosis, in this paper, cost integrated multiclass SVM with reject option (CIMC-SVM) is proposed. Experimental results show that CIMC-SVM is able to improve the diagnosis reliability and minimize the average cost.