最小二乘双支持向量机的在线学习算法

摘要

An online learning algorithm was proposed for solving the problem in designing least square twin sup port vector machines which have two nonparallel planes. By using the lemma of the inverse matrix decomposition, the proposed algorithm can fully utilize the historical training results and avoid the process of calculating the inverse matrix of large matrix, thus reducing the computational complexity. The simulation results demonstrate the effectiveness of the proposed algorithm.%针对具有两个非并行分类超平面的最小二乘双支持向量机,提出了一种在线学习算法.通过利用矩阵求逆分解引理,所提在线学习算法能充分利用历史的训练结果,避免了大型矩阵的求逆计算过程,从而降低了计算的复杂性.仿真结果验证了所提学习算法的有效性.