Automunge

Latest version: v8.33

Safety actively analyzes 723158 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 40 of 99

5.80

- update to __init__ file to simplify import procedure
- now can import AutoMunge directly instead of having to access from Automunger.py
- now imports recomended as
from Automunge import AutoMunge
am = AutoMunge()
- previous version of imports still work as well

5.79

- for oversampling, small update to levelizer function
- added support for oversampling in numeric labels
- based on supplemented ordinal encoded bins
- (previously this was only supported for supplemented one-hot encoded bins)

5.78

- found an opportunity to speed things up a tad
- many instances of searching in list
- now replaced with searching in set
- which runs much faster
- relatively speaking

5.77

- small efficiency improvement to postmunge
- now columns are excluded from assembling NArows support columns
- when not needed for assigned infill types
- (the aggregated internal masterNArows dataframe is seperate from support columns appended to train set)
- this will speed up postmunge a little when not running ML infill / reduce memory overhead

5.76

- important update
- new defaults for two automunge(.) parameters
- MLinfill = True, NArw_marker = True
- based on findings from paper Missing Data Infill with Automunge
- (new paper revision following later today)
- resulting in ML infill on by default
- supplemented by NArw support columns with boolean integers signalling presence of infill
- settings are expected to improve downstream model performance in presence of missing data

5.75

- new MLinfilltype option 'ordlexclude'
- intended for hashed encodings
- 'ordlexclude' is excluded from infill
- purpose is to include hashed sets in ordinal set of columntype_report
- also updated MLinfilltype for hs10 (binary encoded hashed sets) to 'boolexclude'

Page 40 of 99

Links

Releases

Has known vulnerabilities

© 2025 Safety CLI Cybersecurity Inc. All Rights Reserved.