This is a python implementation of the iCMFS + IGFSS (using correlation coefficient) algorithm for feature selection (papers included in the repository). Currently the implementation has been completed on a toy dataset and the 20NG dataset.
See report.pdf for an explanation and the results of our experimentation.
To contribute:
- Fork this repo.
- Clone your forked repo via
git clone <repo URL>. - Create new branch via
git checkout -b branch-name. - Stage files via
git add file-name. - Commit via
git commit -m "message". - Push via
git push -u origin branch-name.
Current Dependencies:
- Scikit-learn
- Numpy
- Run the notebook via Jupyter notebook