The Research on the Method of Feature Selection in Support Vector Machine Based Entropy
The large rotating machinery functioning of the rotor is one of the most important issues. It has great significance to identify the fault early and implement intelligent fault diagnosis. However there is a big nonlinear about large rotating machinery and has less fault samples. This led great difficulties for feature selection and state recognition. Based on Entropy in feature selection, we extract each intrinsic mode’s function energy as eigenvector and make them for input parameter of the support vector machine (SVM) to fault diagnosis. The experiment shows that this method can classify the fault state, and completed intelligent fault diagnosis.
Hao Zhang, Yang Fu and Zhong Tang
X. Y. Zhu et al., "The Research on the Method of Feature Selection in Support Vector Machine Based Entropy", Advanced Materials Research, Vols. 354-355, pp. 1192-1196, 2012