The Application of Feature Selection in Bank Devices Fault Location Based on Correlation Coefficient and Heuristic Search

• 1、National Laboratory of Switching Technology and telecommunication Networks,Beijing University of Posts and Telecommunications,Beijing 100876

Abstract：In order to reduce the dimension of the feature vector in the bank equipment fault location and improve the classification accuracy and time performance of the prediction model, this paper proposed a feature selection method based on the combination of correlation coefficient and heuristic search. In this method, the feature subset is searched through bidirectional search binding the Sequential Forward Search and Sequential Backward Search. Then the redundant features will be removed by computing the correlation coefficient . At last, the optimal feature subset is selected. The na?ve Bayesian Classifier is trained based the selected feature subset. Tests were did for the selection methods based on the Weka platform and UCI data mining databases. The results showed that the feature selection method proposed did good on dimension reduction and improvements on classification accuracy and time performance. It has practical significance in the actual bank devices fault position.

Keywords： Correlation Coefficient Heuristic Search Feature Selection Fault location