5uperpalo/FIREMAN-project_imputation

FIREMAN-project data imputation repository

Data imputation scripts and notebooks related to FIREMAN project and main FIREMAN-project repository. Reporsitory is a work-in-progress project that is part of FIREMAN project activities. Currently it includes GAIN object oriented impelementation using PyTorch and its integration with Raytune distributed parameter searching framework.

• original implementation
• GAIN paper
• any code review and suggestions are welcome

notes included in the repository:

• data imputation
• deep learning notes
• todo list

Example usage notebooks:

• GAIN example notebook
• GAIN training using Raytune
  • solution supports Raytune paramater search framework, which can be monitored by dashboard and progress reporters

install:

python setup.py install

run tests:

pip install -r test-requirements.txt
python -m pytest

documentation:

Automatic HTML documenation generated by sphinx is viewable at Doc and version viewable on github using BitBucket at Doc-git(you have to wait couple of seconds until theme comes up).

generate new documenation (app\docs):

Under docs/:

pip install -r doc-requirements.txt
make html

example dataset

Used in original repository

• UCI Spam

example output available in notebooks