--outliers_to_mean

Switch

--outliers_to_mean [OUTLIER_THRESHOLD]

Description

Set an outlier threshold. After standardization if absolute feature value is greater than threshold then set feature to mean value.

Argument and Default Value

Default threshold is 2.5

Other Switches

Required Switches:

-d, -g, -t

Optional Switches:

Some regression command: --nfold_test_regression, --predict_regression, --test_regression, etc.
Some classification command: --nfold_test_classifiers, --predict_classifiers, --test_classifiers, etc.

Example Commands

# Runs 10-fold cross validation on predicting the users ages from 1grams.
# Set outliers to the default value of 2.5
dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1gram$msgs$user_id$16to16' --outcome_table blog_outcomes \
--outcomes age --combo_test_regression --model ridgecv --folds 10 --outliers_to_mean

# Set the threshold to 3.5
dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1gram$msgs$user_id$16to16' --outcome_table blog_outcomes \
--outcomes age --combo_test_regression --model ridgecv --folds 10 --outliers_to_mean 3.5