Bearings constitute a crucial part of machinery that need to be continuously monitored. Major breakdowns can be prevented if bearing defects are identified at the earlier stage. Sound signals of the bearings can be used to continuously monitor bearing life. This paper uses sound signals acquired in bearings under healthy and simulated faulty conditions for the purpose of fault diagnosis through machine learning approach. The statistical features were extracted from the sound signals. Significantly important features were selected using J48 decision tree algorithm. Support Vector Machine (SVM) is used as a classifier. The selected features were given as inputs for the c-SVM and v-SVM (nu - SVM) model of SVM and their classification accuracies were compared. © Vibroengineering.