--all_controls_only

Switch

--all_controls_only

Description

Works similarly to --control_combo_sizes, but tells the classification or regression predictor to only train using all controls, so giving this switch is equivalent to

--control_combo_sizes N

where N is the number of controls specified with --outcome_controls.

Argument and Default Value

None

Details

Other Switches

Required Switches:

• --outcome_controls

• One of these: --nfold_test_classifiers OR --nfold_test_regression

Optional Switches:

• None

Example Commands

# Only runs including all controls (and no controls, which is done by default):
dlatkInterface.py -d tester7 -t statuses_er1 -c study_code --group_freq_thresh 500 -f 'feat$cat_statuses_er1_cp_w$statuses_er1$study_code$16to16' --outcome_table outcomesFinal_femOnly --outcomes PREGNANCY --controls isWhite isBlack isHispanic ageTercile0 ageTercile1 ageTercile2 age --combo_test_classif --all_controls_only --model lr --folds 10 --csv --output_name ./output