--factor_selection_type

Switch

--factor_selection_type TYPE

Description

Method used to select adaptation factors for --factor_adaptation/--res_factor_adaptation: pca or rfe.

Argument and Default Value

A string. Default: "rfe".

Details

Known issue — the default (``rfe``) is broken. RegressionPredictor.selectAdaptationFactors() calls RFE(model, nFactors) positionally, but the installed scikit-learn (1.6.1) requires n_features_to_select as a keyword-only argument. Verified: leaving this at its default rfe crashes with TypeError: __init__() takes 2 positional arguments but 3 were given as soon as factor selection runs. Workaround, verified working: pass --factor_selection_type pca explicitly — this avoids the broken RFE call entirely and completes normally.

Other Switches

Required Switches:

--factor_adaptation or --res_factor_adaptation

Example Commands

Using the packaged dla_tutorial dataset (see Differential Language Analysis (DLA) Tutorial) — use pca, not the broken default:

dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1to3gram$msgs$user_id$0_05' \
--outcome_table blog_outcomes --outcomes age --outcome_controls gender is_student \
--combo_test_regression --factor_adaptation --adaptation_factors gender is_student \
--num_of_factors 1 --factor_selection_type pca