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/xorigex

Quantlet: xorigst
Description: Quantlet to cut off starting values from matrices of dependent and independent variables for time series analysis. This is needed if one always wants to cut off the same number of observations when analysing AR(p) models with different orders.

Usage: { xnew,ynew } = xorigst(xorig,yorig,xraw,lagmax)
Input:
xorig (nn x p) matrix of lags
yorig (nn x 1) matrix of dependent variable
xraw (n x 1) vector of time series
lagmax scalar, largest lag for consideration (for which starting values have to be cut off)
Output:
xnew ((n-lagmax) x p) matrix of lags
ynew ((n-lagmax x 1) vector of matrix of dependent variable

Example:
pathcafpe = "tp/cafpe/"
func(pathcafpe + "xorigex.xpl")
func(pathcafpe + "xorigst.xpl")
xraw    = #(1:100)
lags    = 1|10
lagmax  = 15
{ xorigau,yorig }= xorigex(xraw,0,lags,"no")
{ xnew,ynew } = xorigst(xorigau[1:rows(yorig),],yorig,xraw,lagmax)
"xnew~ynew" 
xnew~ynew
Result:
Matrix of dependent variable ynew with lagmax observations
cut off such that AR(p) models can be fitted up to
p=lagmax.
The matrix of lagged variables xnew is adjusted accordingly.

Library: cafpe
See also: tp/cafpe/xorigex

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