基于Chebyshev多项式的自适应偏最小二乘回归建模

Construction of Self-adaptive Partial Least Squares Regression Model Based on Chebyshev's Polynomial

把 Chebyshev多项式和偏最小二乘法有机地结合起来 ,提出一种基于 Chebyshev多项式的自适应偏最小二乘回归建模算法 .它自动建立一个预测误差较小的基于Chebyshev多项式的多元多项式模型 ,它的预测效果优于传统多元多项式模型 .本文算法是偏最小乘法的推广 。

This paper provides an algorithm of constructing self adaptive partial least sqaures regression model based on Chebyshev's polynomial.It automatically builds a multinomial model based on Chebyshev's polynomial with less predicting error.Its predicting efficency is better than that of traditional multinomial model.The algorithm proposed in this paper has generalized the partial least square method.In other words it means the application of partial least square method in nonlinear field.