Diagnosis Accuracy in Electric Power Apparatus Conditions using the Classification Methods

摘要

To diagnose the electric power apparatus, the use of the decision tree method was recommended as a classification tool because it provides the if-then-rule in visible, and thus we may have a possibility to connect the physical phenomena to the observed signals. Using a variety of feature variables extracted from the partial discharge patterns and etc., the misclassification rates are found to be very small such as 2% if the results were obtained by using the training data only. We, here, have assessed the diagnosing accuracy of classification methods using the test data. As a result, the small values of the misclassification rates remain even if the test data are applied to the classifiers. The appropriate methods perform fairly well preserving the misclassification rates of less than 5%. The paper concludes that even if the misclassification rates by the decision tree are not so small comparing to attainable values given by the effective classifiers such as bagging, the decision tree is still useful and attractive because the method provide us explicit rules and the variability of the misclassification rates are not so large.