对零售行业的信用评价属性约简进行了研究。利用考虑删去某个条件属性后条件属性与决策属性的互信息是否改变来计算条件属性的属性核。为保障种群的多样性,利用混沌算法对量子蛙的叠加态的观测态赋值及二进制编码,对量子蛙进行适应度评价。按适应度大小对青蛙进行种群分配,对每个种群中最劣青蛙对最优青蛙进行改进步长的量子旋转门学习。经过一定迭代次数后,将所有种群的青蛙进行混合之后学习,再按适应度对青蛙进行种群分配。经过一定种群内学习及种群间学习后使其最优青蛙与最劣青蛙的适应度值之差小于阈值停止迭代从而达到属性约简的目的,从而得出零售行业信用评价的最简属性。 Researches on attribute reduction of credit rating in retail industry were conducted. The attribute kernel of the conditional attribute is calculated by considering whether the mutual information between the conditional attribute and the decision attribute after deleting a certain conditional attribute is changed. In order to ensure the diversity of the population, the chaos algorithm was used to evaluate the superposition state of quantum frogs and binary coding. The frogs were distributed according to their fitness, and the QWRs of the best frogs in each population were studied by the quantum rotation gate. After a certain number of iterations, all the groups of frogs were mixed to learn, and then to the frog population distribution. After a certain population learning and inter-population learning, the difference between the fitness value of the optimal frog and the worst frog is less than the threshold stop iteration so as to achieve the purpose of attribute reduction. Thus, the simplest attribute of the credit rating of the retail industry is obtained.