A scikit-learn compatible package for id3 decision tree
decision-tree-id3 is a module created to derive decision trees using the ID3 algorithm. It is written to be compatible with Scikit-learn’s API using the guidelines for Scikit-learn-contrib. It is licensed under the 3-clause BSD license.
- Python (>= 2.7 or >= 3.3)
- NumPy (>= 1.6.1)
- Scikit-learn (>= 0.17)
The package by itself comes with a single estimator Id3Estimator. To install the module:
pip install decision-tree-id3
or clone the project using:
git clone https://github.com/svaante/decision-tree-id3.git cd decision-tree-id3 python setup.py install
If the installation is successful, you should be able to execute the following in Python:
>>> from sklearn.datasets import load_breast_cancer >>> from id3 import Id3Estimator >>> from id3 import export_graphviz >>> bunch = load_breast_cancer() >>> estimator = Id3Estimator() >>> estimator.fit(bunch.data, bunch.target) >>> export_graphviz(estimator.tree_, 'tree.dot', bunch.feature_names)
And to generate a PDF of the decision tree using GraphViz:
dot -Tpdf tree.dot -o tree.pdf
There are a number of different default parameters to control the growth of the tree: - max_depth, the max depth of the tree. - min_samples_split, the minimum number of samples in a split to be considered. - prune, if the tree should be post-pruned to avoid overfitting and cut down on size. - gain_ratio, if the algorithm should use gain ratio when splitting the data. - min_entropy_decrease, the minimum decrease in entropy to consider a split. - is_repeating, repeat the use of features.
For more in depth information see the documentation https://svaante.github.io/decision-tree-id3
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size decision-tree-id3-0.1.2.tar.gz (12.9 kB)||File type Source||Python version None||Upload date||Hashes View|