最小平方误差(LMSE)算法实现，可训练的确定性分类器的迭代算法。用于对训练一个向量，使得向量与给定矩阵的乘积的结果向量足够小。-least square error (LMSE) algorithm can be trained classifier uncertainty iterative algorithm. For a pair of training vectors, making vector and matrix given the results of the product