提出一种利用最小交叉熵优化方法的矢量量化分类算法， 将矢量的分类问题转化为最小失真函数估计问题，通过关联随机迭代算法逼近最优值，以更好的减小量化误差。-A minimum cross-entropy optimization of the use of methods of classification of vector quantization algorithm, the classification of the vector into the issue of minimum-distortion function is estimated through a random link approximation optimal value iteration algorithm in order to better reduce the quantization error.