Current implementation of SPORF

The current implementation of SPORF is located on the main branch of the repository.

USAGE

The usage of SPORF is exactly the same as scikit-learn's Random Forest algorithm, but with two extra parameters. These parameters are implemented in the same way as RerF

1. max_features is a float (0.0, 1.0) that controls the output dimension of SPORF after applying the projection matrix. Specifically, the output dimension is equal to int(max_features * n_features), where n_features is the number of features of the data.
2. feature_combinations is a float (0.0, n_features) that controls the average number of features to consider at a split.

To build and run SPORF, simply use the following commands:

from oblique_forests.sporf import ObliqueForestClassifier

clf = ObliqueForestClassifier(**params)

clf.fit(X_train, y_train)

y_predict = clf.predict(X_test)

IMPLEMENTATION