【摘要】 介绍了PLS2回归计算顾客满意度指数的建模方法.对于算法中涉及的求高阶实对称阵的最大特征值及其特征向量,采用幂法加以实现.给出了一个具体例子,计算出了顾客满意度、忠诚度等二级指标的值,并对此做了简要分析.PLS2回归较好地克服了各指标间的多重共线性问题,通过此方法求得的顾客满意度指数更准确、合理.更多还原

【Abstract】 The method of model building for the computation of customer satisfaction indices by PLS2 regression is introduced.As for computing the largest eigenvalue and its eigenvector of a higher-order real symmetric matrix in the algorithm,the power method is used.And an example is given.Then the values of level two indices customer satisfaction,customer loyalty are computed and briefly analyzed.PLS2 regression successfully solve the multicollinearity problems among the indices,which indicates that the customer satisfaction indices computed are better and more reasonable.更多还原