--IDP

Switch

--IDP, --idp

Description

Uses Informative Dirichlet Prior (see references) in correlations with binary outcomes instead of linear regression.

Argument and Default Value

No argument. Boolean switch, default: False.

Details

When set, --correlate computes group-level associations with oa.IDP_correlate instead of the standard Pearson/Spearman path. Only one metric switch is used per run: --IDP is checked before --zScoreGroup and --AUC; if none of those three are given, --correlate falls back to its default regression/correlation logic, which separately chooses between --logistic_reg, --cohens_d, --spearman, or Pearson R.

Other Switches

Required Switches: --correlate Optional Switches: None

Example Commands

Since --IDP requires a binary outcome, we use is_student (0/1) from the packaged dla_tutorial dataset (see Differential Language Analysis (DLA) Tutorial):

dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1to3gram$msgs$user_id$0_05' \
--group_freq_thresh 500 --outcome_table blog_outcomes --outcomes is_student \
--correlate --IDP --rmatrix --sort --output_name ./idp_output