Lasso for selection of variables and running logit regression using selected variables in lasso

[Ernest Charles]

Hi all,
Can someone help me please,
I am assessing the impact of income(X1) on economic growth (Y) by using Logit regression approach. Income is my interest variable while i have also control variables (X2, X3, .....X9).
I was told to use lasso to select my variables to use in the regression. Can anyone help me with the command here for my case to select variables using lasso, and command to see the selected variables and finally how can i put the selected variables in my Logit approach to run the regression?

Thank you

[Andrew Musau]

webuse lbw, clear
set seed 08052022
lasso logit low c.age c.lwt i.race i.smoke i.ptl i.ht i.ui i.ftv, nolog
*DISPLAY POST SELECTION VARIABLES
di "`e(post_sel_vars)'"
*LOGIT WITH POST SELECTION VARIABLES
logit `e(post_sel_vars)'

Res.:

. lasso logit low c.age c.lwt i.race i.smoke i.ptl i.ht i.ui i.ftv, nolog
note: 3.ptl dropped because it is constant in C.V. subsamples
note: 6.ftv dropped because it is constant in C.V. subsamples

Lasso logit model                           No. of obs        =        189
                                            No. of covariates =         19
Selection: Cross-validation                 No. of CV folds   =         10

--------------------------------------------------------------------------
         |                                No. of      Out-of-
         |                               nonzero       sample      CV mean
      ID |     Description      lambda     coef.   dev. ratio     deviance
---------+----------------------------------------------------------------
       1 |    first lambda       .1352         0      -0.0154     1.260739
       5 |   lambda before    .0931881         1       0.0076     1.232177
     * 6 | selected lambda    .0849095         1       0.0076     1.232166
       7 |    lambda after    .0773664         1       0.0040     1.236699
       9 |     last lambda    .0642309         6      -0.0050     1.247864
--------------------------------------------------------------------------
* lambda selected by cross-validation.

. 
. *DISPLAY POST SELECTION VARIABLES

. 
. di "`e(post_sel_vars)'"
low ptl

. 
. *LOGIT WITH POST SELECTION VARIABLES

. 
. logit `e(post_sel_vars)'

Iteration 0:   log likelihood =   -117.336  
Iteration 1:   log likelihood = -113.96986  
Iteration 2:   log likelihood = -113.94631  
Iteration 3:   log likelihood = -113.94631  

Logistic regression                             Number of obs     =        189
                                                LR chi2(1)        =       6.78
                                                Prob > chi2       =     0.0092
Log likelihood = -113.94631                     Pseudo R2         =     0.0289

------------------------------------------------------------------------------
         low |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         ptl |   .8018056   .3171535     2.53   0.011     .1801963    1.423415
       _cons |  -.9641889   .1749607    -5.51   0.000    -1.307106   -.6212722
------------------------------------------------------------------------------

[Ernest Charles]

Thank you so much