Binary tree classification algorithm based on GA and its application in the gear machine fault diagnosis

Based on the traditional binary tree for classification problems classification accuracy is not high and high time complexity, we present a new support vector machine fault diagnosis algorithm of a double optimization based on genetic algorithm. It makes full advantage of genetic algorithm to optimize the characteristic parameters subset and nuclear parameters, which is extracted based on wavelet packet, to obtain key parameters of the best equipment failure and effectively solve for different kinds of fault identification. Combining with the gear fault experimental data simulation experiment, the result shows that this algorithm´s validity in this paper.