--all_control_combinations

Switch

--all_control_combinations

Description

Runs every possible subset of --outcome_controls as its own model in --nfold_test_regression/--nfold_test_classifiers, instead of only the no-controls and all-controls models.

Argument and Default Value

No argument. Sets allcontrolsonly to False; default (flag absent) is True.

Details

The opposite of --all_controls_only (same underlying allcontrolsonly variable, set to False here vs True there — whichever is parsed last on the command line wins if both are given). With many controls, the number of subsets grows combinatorially (2^n); for a controlled subset of sizes instead, use --control_combo_sizes.

Other Switches

Required Switches:

--outcome_controls
One of --nfold_test_classifiers or --nfold_test_regression

Example Commands

dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1gram$msgs$user_id$16to16' \
--outcome_table blog_outcomes --outcomes age --outcome_controls gender is_student \
--combo_test_regression --all_control_combinations