--AUC

Switch

--AUC, --auc

Description

Use logistic regression model area under curve (AUC) for --correlate instead of linear regression/correlation [only works with binary outcome values]. "Curve" refers to receiver operating characteristic curve.

Argument and Default Value

Output value for each predictor feature will range from 0.5 (completely non-predictive) to 1.0 (perfectly predictive).

Details

Other Switches

Required Switches:

Optional Switches:

Example Commands

daltkInterface.py -d tester7 -t statuses_er1 -c study_code --group_freq_thresh 500 \
-f 'feat$cat_ser1_f2_200_cp_w$statuses_er1_655$study_code$16to16' \
--outcome_table outcomesFinal --outcomes DM_UNCOMP \
--correlate --rmatrix --controls sex_int isWhite isBlack isHispanic ageTercile0 ageTercile1 ageTercile2 \
--auc --no_correction --csv --output_name OUTPUT