Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Library: cafpe
See also: tp/cafpe/dencp

Quantlet: fvllc
Description: Quantlet for multivariate local linear or partial local quadratic estimation using C++ - routines via dlls. It can estimate conditional means, conditional volatilities, first derivatives, second direct derivatives, conditional densities with full or leave-one-out and density estimation with all possible data or only lagged data. Only the gaussian kernel can be used.


Reference(s):

Usage: { mA,gsqA,denA,err } = fvllc(Xsj,Yorig,h,Xtj,kernreg,lorq,fandg,loo);
Input:
ATTENTION: this quantlet requires to open locling.dll, density.dll (NT) or locling.so, denc.so (UNIX). This can be done with the quantlet cafpeload or directly with garb = dlopen ("\locling.dll") on NT, garb = dlopen ("/locling.so") on UNIX.
Xsj (n x d) matrix of regressors
Yorig (ny x 1) vector of dependent data
h scalar or (d x 1) vector, bandwidth
Xtj (nxest x d) matrix of data at which estimation is conducted
kernreg must currently be "gaussian"
lorq 1=local linear, 2=local quadratic with direct second derivatives
fandg 0=only mean function, 1=also volatility function (be careful: volatility function may not be positive)
loo 0=take all, 1 = leave-one-out (only works if n=ny)
Output:
mA (nxest x (lorq*d+1)) matrix of cond. mean, cond. first der., cond. direct sec. der. (lorq=2)
gsqA (nxest x (lorq*d+1)) matrix of cond. volatility, cond. first der., cond. direct sec. der. (lorq=2)
denA (nxest x 1) vector of cond. density
err 0 if matrix inversion in dll ok, 1 if matrix inversion failed

Library: cafpe
See also: tp/cafpe/dencp

Keywords - Function groups - @ A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Author: Tschernig 000420
(C) MD*TECH Method and Data Technologies, 27.4.2000