Version 0.7

This release adds the ability to separate in-bag and out-of-bag samples for any forest model. We also introduce a new class for fitting honest forests while permuting the covariate index, and a new set of simulations based on Marron and Wand 1992.

In addition, various patches were made in terms of how to use scikit-tree for hypothesis testing of feature sets.

Changelog

• Feature Introduce a new light-weight class for fitting honest forests while

  permuting the covariate index sktree.stats.PermutationHonestForestClassifier, by Adam Li (##211)

• Feature Introduce a new class method predict_proba_per_tree for all

  Forest classifiers, which will predict the probability per tree and keep the output as a (n_estimators, n_samples, n_classes) output, by Adam Li (##211)

• Feature Introduce a new class fitted attribute oob_samples_ for all

  Forest models, which will keep track of the samples used. by Adam Li (##211)

• Feature Introduce a new set of simulations based on Marron and Wand 1992.

  by Sambit Panda (##203)

• Feature sktree.stats.build_coleman_forest() and sktree.stats.build_permutation_forest()

  are added to compute p-values given an estimator and permutation-estimator, Adam Li (##222)

• API Change sktree.datasets.make_trunk_classification() for generating trunk mixture and Marron-Wand

  simulations are separated out into sktree.datasets.make_marron_wand_classification() and sktree.datasets.make_trunk_mixture_classification(), Adam Li (##227)

• API Change sktree.HonestForestClassifier and sktree.tree.HonestTreeClassifier

  now overwrite all parameters set by the underlying tree_estimator and allow you to directly pass any extra parameters that tree_estimator has compared to the original DecisionTreeClassifier, Adam Li (##228)

• Fix Trunk simulators now correctly generate random values with a fixed seed,

  by Sambit Panda (##236)

• Fix Trunk simulators now correctly generate random values with a fixed seed,

  by Sambit Panda (##236)

• Efficiency All scikit-tree estimators are now at least 2X faster than they were

  in previous versions. This was due to adding in compiler-directives to turn on optimizations ‘-03’ when compiling the C++ generated code from Cython. In addition, we explicitly turned off bounds-checking and related runtime checks in the Cython code, which would lead to performance degradation during runtime. by Adam Li (##242)

Code and Documentation Contributors

Thanks to everyone who has contributed to the maintenance and improvement of the project since version inception, including:

• Adam Li

• Sambit Panda