--test_c2r

Switch

--test_c2r

Description

Trains and tests a "classify-to-regression" model: first classifies groups into bins, then regresses within/across those classifications to predict the original continuous outcome.

Argument and Default Value

No argument. Boolean switch, default: False.

Details

Verified requirement, not fully exercised: ClassifyToRegressionPredictor looks up outcomes using a bin_<outcome> column name (e.g. bin_age for an outcome named age) rather than the plain outcome column — confirmed by running against a table without one and seeing Loading Outcomes and Getting Groups for: {'bin_age'} followed by a "no such column: bin_age" error. You need a pre-binarized version of your outcome in the outcome table under that naming convention (e.g. via --categorical) before this switch can run. Not independently verified past this point due to time constraints in this pass.

Other Switches

Required Switches:

• -d, -g, -t

• -f

• --outcome_table, --outcomes (with a matching bin_<outcome> column present)

Example Commands

Using the packaged dla_tutorial dataset (see Differential Language Analysis (DLA) Tutorial), assuming a bin_age column exists in blog_outcomes:

dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1to3gram$msgs$user_id$0_05' \
--outcome_table blog_outcomes --outcomes age --test_c2r