偏最小二乘法应用于回归的局限性

Limitation of Partial Least Square Method in Regression

通过剖析"偏最小二乘回归模型"的原理,指出其局限性以及使用的必要条件。通过反例证明用"偏最小二乘法"建立回归方程出现不合理结果的可能性。最后通过一个成分数据的实例,验证了"偏最小二乘回归模型"在该类问题中的不适宜性。

Through analyzing the principle of partial least square regression model, its limitations and the necessary conditions for its use are pointed out. Using a counterexample, it is illustrated that irrational result appears when the regression equation is built by the partial least square method. At the end we testify that partial least square regression model does not adapt to the problem through an example of the composition data.