Skip to main content

Decision Forest C++ library with a scikit-learn compatible Python interface

Project description

Travis Codecov ReadTheDocs

koho (TM)

koho (Hawaiian word for ‘to estimate’) is a Decision Forest C++ library with a scikit-learn compatible Python interface.

  • Classification

  • Numerical (dense) data

  • Missing values (Not Missing At Random (NMAR))

  • Class balancing

  • Multi-Class

  • Multi-Output (single model)

  • Build order: depth first

  • Impurity criteria: gini

  • n Decision Trees with soft voting

  • Split a. features: best over k (incl. all) random features

  • Split b. thresholds: 1 random or all thresholds

  • Stop criteria: max depth, (pure, no improvement)

  • Bagging (Bootstrap AGGregatING) with out-of-bag estimates

  • Important Features

  • Export Graph


New BSD License

Change Log: 1.1.0 Multi-Output (single model) 1.0.0 Missing Values (NMAR) : Python, Cython(bindings), C++ 0.0.2 Criterion implemented in Cython 0.0.1 Classification : Python only

Copyright 2019, AI Werkstatt (TM). All rights reserved.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

koho-1.1.0.tar.gz (159.1 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page