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: eiv
See also: eivvec2

Quantlet: eivvec1
Description: eivvect1 presents the maximum likelihood estimators of the parameters in the measurement error models, which has more than one variable x. The covariances between e and u, Sigeu and the covariance matrix of u, siguu are known. All of the variables obey normal distributions. All parameters are estimated by maximum likelihood method in measurement error models.

Reference(s):

Link:
Usage: {mux,hatbeta,beta0,hatsigmae,hatsigmax)=eivvec1(w,y,sigue,siguu)
Input:
w n x p matrix, the design variables
y n x 1 matrix, the response
sigue p x 1 matrix, the vector of covariances between u and e
siguu p x p matrix, the covariance matrix of U
Output:
mux scalar, the mean value of x
hatbeta1 vector, the estimate
hatbeta0 scalar, the estimate
hatsigmax p x p matrix, the estimate of the covariance matrix of x
hatsigmae scalar, the estimate of the variance of error e

Example:

library("xplore")

library("eiv")

n = 100

randomize(n)

nu =#(2,3,4)

sig=0*matrix(3,3)

sig[,1]=#(0.25, 0.9, 0.1)

sig[,2]=#(0.9, 1, 0.2)

sig[,3]=#(0.1, 0.2, 4)

x=normal(n,3)*sig+nu'

w=x+0.01*normal(n,3)

a1=#(1.2, 1.3, 1.4)

y=0.75+x*a1+0.09*normal(n)

sigue=#(0.11, 0.09, 045)

siguu=0*matrix(3,3)

siguu[,1]=#(1.25, 0.009, 0.01)

siguu[,2]=#(0.009,0.081, 0.02)

siguu[,3]=#(0.01, 0.02, 1.96)

gest=eivvec1(w,y,sigue,siguu)

gest.mux

gest.hatbeta

gest.beta0

gest.hatsigmax

gest.hatsigmae

Result:

Contents of mux

[1,]    2.024   2.9106   3.9382 

Contents of hatbeta

[1,]   0.011384  

[2,]   0.013461 

[3,]   0.013913

Contents of beta0

[1,]   12.362  

Contents of hatsigmax

[1,]  0.84466   1.0319  0.43677 

[2,]   1.0319    1.664   1.0941 

[3,]  0.43677   1.0941   19.781 

Contents of hatsigmae

[1,]   1034.9 


Library: eiv
See also: eivvec2

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: Hua Liang, 970725
(C) MD*TECH Method and Data Technologies, 21.9.2000