<< Automatic estimation Automatic estimation auto_stage1 >>

Grocer >> Automatic estimation > auto_stage0

auto_stage0

elimination of non significative variables

CALLING SEQUENCE

[done,r0,indxf]=auto_stage0(y,x,indcte,r00,rmod,alpha,f0_sig,eta,z)

PARAMETERS

Input

• y = vector of the endogenous variable

• x = matrix of the exogenous variables

• indcte = index of the constant variable (= nvar+1 if there is no constant in the r00 model)

• r00 = results of the estimation (by ols or ols2) of the model Y = X*b+U

• rmod = a tlist defined by def_results

• alpha = broad significance level (typically lower than f0_sig)

• f0_sig = maximum joint significance level allowing elimination of the corresponding variables

• eta = (px1) significance level for specification tests

• z = matrix of the compulsory variables (the ones that must be in the regression whatever significance they have)

Output

• done = %t if all variables are significant, %f otherwise

• r0 = results structure of the final model (the one without the eliminated variables) as provided by ols2

• indxf = indexes in x of the remaining exogenous variables

DESCRIPTION

In a model Y=X*b+U, eliminates variables whose joint significance is lower than f0_sig under the condition that no specification test fails. The function (used in automatic) assumes that an empty result tlist exists and that the model Y==X*b+U with tlist results called r00 in the function. The user must give the index, if any, of the constant, the global significance level, the significance level under which elimination is done and the significance level for the specifications tests performed by function test_func.

EXAMPLE

[done,r0,indxf]=auto_stage0(y,x,5,r00,rmod,0.05,0.9,0.01*ones(5,1))
[done,r0,indxf]=auto_stage0(grocer_y,grocer_x,indcte,r1_00,lrmod(1),grocer_alpha, grocer_f0_sig,grocer_eta)

// In example 1, function auto_stage0 is called with a 0.05 general significance level, a 0.9 elimination level and a 0.01 significance level for 5 specification tests. Example 2 is the call to function auto_satge0 in automatic (the most natural use).

AUTHOR

Eric Dubois 2002

<< Automatic estimation Automatic estimation auto_stage1 >>