Package for interpreting scikit-learn's decision tree and random forest predictions. Allows decomposing each prediction into bias and feature contribution components as described in https://blog.datadive.net/interpreting-random-forests/. For a dataset with n features, each prediction on the dataset is decomposed as prediction = bias + feature_1_contribution + ... + feature_n_contribution.

It works on scikit-learn's

• DecisionTreeRegressor
• DecisionTreeClassifier
• ExtraTreeRegressor
• ExtraTreeClassifier
• RandomForestRegressor
• RandomForestClassifier
• ExtraTreesRegressor
• ExtraTreesClassifier

Free software: BSD license

• scikit-learn 0.17+

The easiest way to install the package is via pip:

$ pip install treeinterpreter

from treeinterpreter import treeinterpreter as ti
# fit a scikit-learn's regressor model
rf = RandomForestRegressor()
rf.fit(trainX, trainY)

prediction, bias, contributions = ti.predict(rf, testX)

Prediction is the sum of bias and feature contributions:

assert(numpy.allclose(prediction, bias + np.sum(contributions, axis=1)))
assert(numpy.allclose(rf.predict(testX), bias + np.sum(contributions, axis=1)))

More usage examples at https://blog.datadive.net/random-forest-interpretation-with-scikit-learn/.