An incremental learning algorithm for improved least squares twin support vector machine

摘要

In this paper, we mainly propose an incremental version of improved least squares twin support vector machine (IILSTSVM), based on inverse matrix-free method. This algorithm can meet the requirement of online learning to update the existing model. In the case of low dimension data, this method effectively improves training speed of incremental learning. According to updating inverse matrix, we can implement the incremental learning for ILSTSVM. Experiments prove that this algorithm has excellent performance on runtime and recognition rate in the low dimensional space.