提出误差选择竞争学习算法，它把遗传算法中的选择机制引入到矢量量化设计中，在使用竞争学习算法减小期望误差的前提下，利用选择机制调整各个区域的子误差从而进一步改善期望误差，实验结果表明，该算法较好地调整了各区域的子误差，克服局部最优 In this paper, the error selection competition learning algorithm is proposed, which introduces the selection mechanism of genetic algorithm into the vector quantization design. Using the competition learning algorithm to reduce the expected error, the selection mechanism is used to adjust the sub-error of each region to further improve the expected error. Experimental results show that the proposed algorithm adjusts the sub-errors of each region well and overcomes the local optimum