摘要
Quasi-Regression主要用于解决高维空间的函数逼近问题 .函数的拟合效果依赖于所选择的标准正交基函数和拟合算法 .在很多情况下 ,未知函数可以表示成若干个部分变量的函数和 .本文对此提出了一种函数的拟合算法 ,通过对这些部分变量的函数的拟合 ,得到原函数的拟合 ,并且说明新方法精度更高 ,计算量更少 .
Quasi-regression was introduced for approximation of a function in high dimensional space in recent literature (reference[2]). The accuracy of approximation to a function depends on the choice of standard orthogonal basis functions and fitting algorithm. In many cases, according to experiences and historical data, it can be deduced that an unknown function is equal to the sum of some other functions each of which only separately includes some of the variables. In this paper an approach for approximating an unknown function from data is given. By approximating these functions respectively, a very good approximation of an unknown function was abtained. It is also verified that the new approach is both more accurate and more computationally efficient than usual ones.
出处
《南开大学学报(自然科学版)》
CAS
CSCD
北大核心
2003年第1期44-49,共6页
Acta Scientiarum Naturalium Universitatis Nankaiensis
基金
国家自然科学基金资助项目 (1 0 1 71 0 5 1 )
高等学校博士学科点专项科研基金资助项目 (1 9990 0 5 5 1 2 )