<< des_stat Econometric tests and diagnostics diebmar >>

Grocer >> Econometric tests and diagnostics > dfbeta

dfbeta

BKW influential observation diagnostics

CALLING SEQUENCE

[results]=dfbeta(y, arg1,...,argn)

PARAMETERS

Input

* y = either an ols results tlist or a a time series, a real (nx1) vector or a string equal to the name of a time series or a (nx1) real vector between quotes

* argi = an argument which can be:

  - the string 'noprint' if the user doesn't want to print the results of the regression

or, and only if y is not an ols results tlist:

  - a time series

  - a real (nx1) vector

  - a string equal to the name of a time series or a (nx1) real vector between quotes

  - 'dropna' if the user wants to remove the NA values from the data

 

Output

* results = a results tlist with

  - results('meth') = 'dfbeta'

  - results('nobs') = # of observations

  - results('nvar') = # of variables in x-matrix

  - results('dfbeta') = df betas

  - results('dffits') = df fits

  - results('hatdi') = hat-matrix diagonals

  - results('stud') = studentized residuals

  - result('namex') = name of the x variables

  - result('namey') = name of the y variable

  - result('prests') = boolean indicating the presence or absence of a time series in the regression

  - result('dropna') = boolean indicating if NAs have been droped

  - result('nonna') = vector indicating position of non-NAs (if the option 'dropna' was active)

DESCRIPTION

Computes BKW (influential observation diagnostics) dfbetas, dffits, hat-matrix, studentized residuals. Plots the corresponding results.

EXAMPLE

exo = grand(100,6,'nor',0,1);
exo(:,1) = ones(100,1);
exo(:,3) = exo(:,2) + grand(100,1,'nor',0,1)*0.05;
 
bet = ones(6,1);
 
endo = exo*bet + grand(100,1,'nor',0,1);
 
// now add a few outliers
endo(50,1) = 10.0;
endo(70,1) = -10.0;
 
result = dfbeta('endo','exo');
// This example is taken from function dfbeta_d. The example provides dfbetas, dffits,
// hat-matrix, studentized residuals for a regression whose endogenous variable is y
// and exogenous variables is a matrix (as in dfbeta_d), a vector or a ts x.

AUTHOR

Eric Dubois 2002

Report an issue
<< des_stat Econometric tests and diagnostics diebmar >>