Batch GMM estimation
Posted: Wed Sep 28, 2011 4:21 pm
I would like to like to run GMM estimations on several hundred series, one at a time. For instance, I would be using a command similar to
gmmc.gmm(instwgt=white,gmmiter=2,nodf) cons c1 y1 y1(-1) w1 @ c1 p1(-1) k1(-1) x1(-1)
gmmc.gmm(instwgt=white,gmmiter=2,nodf) cons c2 y2 y2(-1) w2 @ c2 p2(-1) k2(-1) x2(-1)
...
gmmc.gmm(instwgt=white,gmmiter=2,nodf) cons c300 y300 y300(-1) w300 @ c300 p300(-1) k300(-1) x300(-1)
for the estimation. Now instead of just displaying the estimation output in the equation window I am looking for a way how I can address the individual elements of the equation output (such as coefficients, standard errors, t-stats and associated probabilities) directly from the command line. I would like to implement a batch run for each of the several hundred series and save estimation results in a consolidated format either within an Eviews table or by exporting it to an Excel/CSV file.
How can I address the estimation results from the command line and what would be the best way of aggregating the results for each of the several hundred separate estimations in a consolidated table?
Thank you very much.
gmmc.gmm(instwgt=white,gmmiter=2,nodf) cons c1 y1 y1(-1) w1 @ c1 p1(-1) k1(-1) x1(-1)
gmmc.gmm(instwgt=white,gmmiter=2,nodf) cons c2 y2 y2(-1) w2 @ c2 p2(-1) k2(-1) x2(-1)
...
gmmc.gmm(instwgt=white,gmmiter=2,nodf) cons c300 y300 y300(-1) w300 @ c300 p300(-1) k300(-1) x300(-1)
for the estimation. Now instead of just displaying the estimation output in the equation window I am looking for a way how I can address the individual elements of the equation output (such as coefficients, standard errors, t-stats and associated probabilities) directly from the command line. I would like to implement a batch run for each of the several hundred series and save estimation results in a consolidated format either within an Eviews table or by exporting it to an Excel/CSV file.
How can I address the estimation results from the command line and what would be the best way of aggregating the results for each of the several hundred separate estimations in a consolidated table?
Thank you very much.