Support Vector Machine Ensemble Based on Feature and Hyperparameter Variation for Real-World Machine Fault Diagnosis

The support vector machine (SVM) classifier is currently one of the most powerful techniques for solving binary classification problems. To further increase the accuracy of an individual SVM we use an ensemble of SVMs, composed of classifiers that are as accurate and divergent as possible. We investigate the usefulness of SVM ensembles in which the classifiers differ among themselves in both the feature set and the SVM parameter value they use, which might increase the diversity among the classifiers and therefore the ensemble accuracy. We propose a novel method for building an accurate SVM ensemble. First we perform complementary feature selection methods to generate a set of feature subsets, and then for each feature subset we build a SVM classifier which uses tuned SVM parameters. The experiments show that this method achieved a higher estimated prediction accuracy in comparison to well-established approaches for building SVM ensembles, namely using a Genetic Algorithm based search to vary the classifier feature sets and using a predefined set of SVM parameter values to vary the classifier parameters. We work in a context of real-world industrial machine fault diagnosis, using 2000 examples of vibrational signals obtained from operating faulty motor pumps installed on oil platforms.