Regress::PCRegress Function

Overload List

#SignatureDescription
1void PCRegress(TVec *Y, TMtx *A, TVec *b, TVec *YCalc = null, TVec *Bse = null, int NumOmmit = 1);Principal Component Regression.
2void PCRegress(TVec *Y, TMtx *A, TVec *b, TVec *Weights, TVec *YCalc = null, TVec *Bse = null, int NumOmmit = 1);Weighted PC regression.

Overload 1: void PCRegress(TVec *Y, TMtx *A, TVec *b, TVec *YCalc = null, TVec *Bse = null, int NumOmmit = 1);

Principal Component Regression.

#NameTypeDescription
1YTVec *Defines vector of dependant variable.
2ATMtx *Defines matrix of independant variables.
3bTVec *Returns calculated regression coefficiens.
4YCalc = nullTVec *Returns vector of calculated dependant variable, where YCalc = A*b + constant term.
5Bse = nullTVec *Returns principal component b coefficient standard error.
6NumOmmit = 1intDefines the number of variables to ommit from initial model.
Remarks:

Performs unweighted Principal Component Regression (PCR). PCR is a technique for analyzing multiple regression data that suffer from multicollinearity. When multicollinearity occurs, least squares estimates are unbiased, but their variances are large so they may be far from the true value. By adding a degree of bias to the regression estimates, principal components regression reduces the standard errors. The algorithm first standardizes A matrix and performs PC regression on standardized matrix.

See Also: Regress::RidgeRegress
Declared in Dew::Stats::Units::Regress · Dew.Stats/Units.Regress.h · Cross-compiler

Overload 2: void PCRegress(TVec *Y, TMtx *A, TVec *b, TVec *Weights, TVec *YCalc = null, TVec *Bse = null, int NumOmmit = 1);

Weighted PC regression.

#NameTypeDescription
1YTVec *Defines vector of dependant variable.
2ATMtx *Defines matrix of independant variables.
3bTVec *Returns calculated regression coefficiens.
4WeightsTVec *Defines weights for PC regression.
5YCalc = nullTVec *Returns vector of calculated dependant variable, where YCalc = A*b + constant term.
6Bse = nullTVec *Returns principal component b coefficient standard error.
7NumOmmit = 1intDefines the number of variables to ommit from initial model.
Declared in Dew::Stats::Units::Regress · Dew.Stats/Units.Regress.h · Cross-compiler