Skip to content

Conversation

@moprescu
Copy link
Contributor

Modified docs, README and added a new notebook

}
],
"source": [
"plt.plot(X_test, test_effect, \"--\", label=\"Truth\")\n",
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Weird that the final approach that does a CV among both lasso and forest is worse than just lasso. Doesn't seem right.

"\n",
"An alternative to passing in CV models to EconML is to perform model selection outside of the EconML estimators and then pass in the pre-selected models to EconML.\n",
"\n",
"**Advantages:** Faster runtimes of the EconML estimators and more flexible selection of first stage models.\n",
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Add here also that as long as log(number of hyper-parameters) is much smaller than the number of samples, this approach maintains statisical validity of the resulting inference results.

Miruna Oprescu added 2 commits January 15, 2021 12:16
* Modified docs, README and added a new notebook
@moprescu moprescu force-pushed the moprescu/nuisance_cv branch from e32171f to eeaf2a4 Compare January 15, 2021 17:16
@moprescu moprescu merged commit fb34846 into master Jan 15, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants