An incremental algorithm of support vector machine based on distance ratio and k nearest neighbor

Abstract

For large data sets and data updated situation, incremental training algorithm is an effective solution of support vector machine training. To improve speed of incremental support vector machine training algorithm, this paper combines the distance ratio method and the nearest neighbor method to extract boundary samples, and an incremental support vector machine algorithm based on distance ratio and k nearest neighbor was proposed, this algorithm can eliminate useless samples as far as possible, thus reduces the training time at remaining essentially the same training accuracy.