Implementation of a rule based prediction algorithm called RIPE (Rule Induction Partitioning Estimate). RIPE is a deterministic and interpretable algorithm, for regression problem. It has been presented at the International Conference on Machine Learning and Data Mining in Pattern Recognition 2018 (MLDM 18). The paper is available in arXiv https://arxiv.org/abs/1807.04602.
These instructions will get you a copy of the project up and running on your local machine for development and testing purposes. See deployment for notes on how to deploy the project on a live system.
RIPE is developed in Python version 2.7. It requires some usual packages
- NumPy (post 1.13.0)
- Scikit-Learn (post 0.19.0)
- Pandas (post 0.16.0)
- SciPy (post 1.0.0)
- Matplotlib (post 2.0.2)
- Seaborn (post 0.8.1)
See requirements.txt.
sudo pip install package_name
To install a specific version
sudo pip install package_name==version
The latest version can be installed from the master branch using pip:
pip install git+git://github.com/VMargot/RIPE.git
Another option is to clone the repository and install using python setup.py install or python setup.py develop.
RIPE has been developed to be used as a regressor from the package scikit-learn.
from sklearn import datasets
iris = datasets.load_iris()
X, y = iris.data, iris.target
ripe = RIPE.Learning()
ripe.fit(X, y)
ripe.predict(X)
ripe.score(X,y)
To have the Pandas DataFrame of the selected rules
ripe.selected_rs.to_df()
Or, one can use
ripe.make_selected_df()
To draw the distance between selected rules
ripe.plot_dist()
To draw the count of occurrence of variables in the selected rules
ripe.plot_counter_variables()
This implementation is in progress. If you find a bug, or something witch could be improve don't hesitate to contact me.
- Vincent Margot
See also the list of contributors who participated in this project.
This project is licensed under the GNU v3.0 - see the LICENSE.md file for details