Speeding up multi-SVMs through modified working set selection

Abstract

Support vector machines (SVMs) possess good accuracy in big data classification. However, the computational cost in both training and testing stages is a critical issue. In this study, we speed up the training speed through an integrated working set selection which is based on a modification of the selection pool of the working set of SVMs. In the testing stage, a lost-min strategy is proposed to accelerate the voting algorithm used in multi-SVMs. The number of the used binary classifiers is reduced from an order of to (nearly to or ). The proposed methods were tested with DNA dataset (bioinformatics), Usps datasets (handwritten digits), Letter dataset (English alphabet) and Satimage dataset (satellite imagery of Earth). We further theoretically derive the time complexity of the proposed method approaches to algorithm in the case of high accuracy. This result is demonstrated through the experimental results for these datasets.