天软金融分析.NET函数大全 > TSL函数 > 数学函数 > 回归 > 回归检验

Regress_Ridge    

简述
岭回归，程序中对x进行了标准化，再结合岭回归系数去计算回归方程的系数，程序返回的结果，每一行为一组回归系数。
Regress_Ridge(y,x)
输出x对y的岭回归，k的选取依据方差扩大因子法
Regress_Ridge(y,x,k)
k是一个实数，输出x对y的岭回归，k是一个长度大于1的数组，输出x对y的岭迹
定义
Regress_Ridge(y:Array of Real;x:Array of Real;k:Array of Real;constant:bool;alpha:real):array;
参数
名称类型说明
YArray of Real 被解释变量序列，为一维数组类型
XArray of Real 解释变量矩阵，为二维数组类型，每一列为一个解释变量
kArray of Real 岭回归系数，0到1之间的实数，可为整数或一维数组类型；缺省使用方差扩大因子选取
Constantbool是否包含常数项，为布尔类型，缺省为1
Alphareal显著性水平，实数或整数，在01之间
  • 范例

    y:=array(49.00,50.2,50.5,48.5,47.5,44.5,28.00,31.5,34.5,35.00,38.00,38.5,15.00,17.00,20.5,29.5);
    x:=array(
      (1300.00,7.5,0.012,9750.00,15.6,0.09),
      (1300.00,9.00,0.012,11700.00,15.6,0.108),
      (1300.00,11.00,0.0115,14300.00,14.95,0.1265),
      (1300.00,13.5,0.013,17550.00,16.9,0.1755),
      (1300.00,17.00,0.0135,22100.00,17.55,0.2295),
      (1300.00,23.00,0.012,29900.00,15.6,0.276),
      (1200.00,5.3,0.04,6360.00,48.00,0.212),
      (1200.00,7.5,0.038,9000.00,45.6,0.285),
      (1200.00,11.00,0.032,13200.00,38.4,0.352),
      (1200.00,13.5,0.026,16200.00,31.2,0.351),
      (1200.00,17.00,0.034,20400.00,40.8,0.578),
      (1200.00,23.00,0.041,27600.00,49.2,0.943),
      (1100.00,5.3,0.084,5830.00,92.4,0.4452),
      (1100.00,7.5,0.098,8250.00,107.8,0.735),
      (1100.00,11.00,0.092,12100.00,101.2,1.012),
      (1100.00,17.00,0.086,18700.00,94.6,1.462));
    k:=array();
    i:=0;
    for nI:= 0 to 6e-3 step 1e-5 do
    k[i++]:=nI;
    return ret:= Regress_Ridge(y,x,k,false);//输出岭迹
    // return ret:= Regress_Ridge(y,x);// 输出x对y的岭回归，方差扩大因子法选取k


    参考
    Regression Regress_pri Regress_Stepwise Regress_VIF boxcox 
相关