回归函数的有理逼近模型分析与研究

经济问题的研究中存在着大量的回归分析问题 ,但变量之间的关系往往是非线性的 .传统的建模原则往往对问题作了一系列的假设 ,因而模型不具有普遍适应性 .本文引进了一种特殊的非参数估计方法——回归函数的有理逼近 ,它与最小二乘法相比 ,提高了拟合与预测的精度 .

Analysis and Research about the Model of the Rational Approximation of Regression Function

Although there are a lot of problems about regressive analysis in the research of economic issues, the relations among variables are usually nonlinear. The traditional modeling principle often adds a series of hypotheses for solving these problems, therefore these models haven′t general adaptabilities. This paper introduces a special method of nonparametric estimation—the rational approximation of aggressive function, which, comparing with the least quadratic multiplication, enhances the precision of between fit and forecast.