hyppo.independence.sim_matrix(model, x)

Computes the similarity matrix from a random forest.

The model used must follow the scikit-learn API. That is, the model is a random forest class and is already trained (use fit) before running this function. Also, func:apply must be used to push input data down the trained forest. See sklearn.ensemble.RandomForestClassifier and sklearn.ensemble.RandomForestRegressor for similar classes.

  • model (class) -- A trained random forest object for which the similarity matrix will be calculated.

  • x (ndarray) -- Input data matrice. x must have shape must be (n, p) where n is the number of samples and p are the number of dimensions.


proxMat (ndarray) -- The proximity matrix induced by random kernel.