“…Parameters (C, r) of support vector machine are selected from [0.5, 50] and 10-fold cross validation method is used. Ant colony algorithm [38,39] is also adopted for parameters optimization of C and r. intelligent fault diagnosis method with multivariable ensemble-based incremental support vector machine, which are used in the following Tables. From Table 5, it is seen that the proposed method and typical support vector machine exhibit better performance than other seven methods (Discrete Cosine Transform, Daubechies wavelet, Symlets wavelet, Walsh transform, FFT, Walsh-Rough set theory, FFT-Rough set theory) in identifying three common fault conditions (ball fault, inner race fault and outer race fault), which hit the highest accuracy of 100%.…”