Modified Kernel Marginal Fisher Analysis for Feature Extraction and Its Application to Bearing Fault Diagnosis

Table 5

The average recognition rates (%) of the simple features based on KNN.

Case  1: fault categories identification					Case  2: fault severities identification
Training sample size of each class
10	30	50	70	90
68.89	94.29	98.33	99.67	99.95	53.25