adaXT

Fast Adaptable and Extendable Trees for Research

adaXT is a Python module for tree-based machine-learning algorithms that is fast, adaptable and extendable. It aims to provide researchers a more flexible workflow when developing tree-based models.

It is distributed under the 3-Clause BSD license.

We encourage users and developers to report problems, request features, ask for help, or leave general comments.

Website: https://NiklasPfister.github.io/adaXT

Overview

adaXT implements several tree types that can be used out-of-the-box, both as decision trees and random forests. Currently, the following tree types are implemented:

Classification: For prediction tasks in which the response is categorical.
Regression: For prediction tasks in which the response is continuous.
Quantile: For uncertainty quantification tasks in which the response is continuous and the goal is to estimate one or more quantiles of the conditional distribution of the response given the predictors.
Gradient: For tasks in which one aims to estimate (directional) derivatives of the response given the predictors. A related tree type is used in the Xtrapolation method.

Beyond these pre-defined tree types, adaXT offers a simple interface to extend or modify most components of the tree models. For example, it is easy to create a custom criteria function that is used to create splits.

Getting started

adaXT is available on pypi and can be installed via pip

pip install adaXT

Working with any of the default tree types uses the same class-style interface as other popular machine learning packages. The following code illustrates this for Regression and Quantile random forests:

from adaXT.random_forest import RandomForest
import numpy as np

# Create toy regression data
n = 100
X = np.random.normal(0, 1, (n, 2))
Y = X[:, 0] + np.random.normal(0, 1, n)
Xtest = np.c_[np.linspace(-1, 1, n), np.random.uniform(0, 1, n)]

# Task 1: Fit regression forest
rf = RandomForest("Regression")
rf.fit(X, Y)

# Predict on test data
Ypred = rf.predict(Xtest)

# Predict forest weight on X or Xtest
# -- can be used a similarity measure on the predictor space
weight_train = rf.predict_weights()
weight_test = rf.predict_weights(Xtest)

# Task 2: Fit a quantile regression
qf = RandomForest("Quantile")
qf.fit(X, Y)

# Predict 10% and 90% conditional quantile on test data
Ybdd = qf.predict(Xtest, quantile=[0.1, 0.9])

The main advantage of adaXT over existing tree-based ML packages is its modularity and extendability, which is discussed in detail in the documentation.

Project goals

This project aims to provide a flexible and unified code-base for various tree-based algorithms that strikes a balance between speed and ease with which the code can be adapted and extended. It should provide researchers a simple toolkit for prototyping new tree-based algorithms.

adaXT provides an intuitive user experience that is similar to the scikit-learn implementation of decision trees both in terms of model-based syntax and hyperparameters. Under the hood, however, adaXT strikes a different balance between speed and ease of adapting and extending the code.

Adaptable and extendable

At the heart of any tree-based algorithm is a decision tree that can be fitted on data and then used to perform some version of prediction. adaXT has therefore been designed with a modular decision tree implementation that takes four input components:

Criteria class: Used during fitting to determine splits.
LeafBuilder class: Used during fitting to specify what is saved on the leaf nodes.
Splitter class: Used during fitting to perform the splits.
Predict class: Used after fitting to make predictions.

By specifying these four components a range of different tree algorithms can be created, e.g., regression trees, classification trees, quantile regression trees and gradient trees. Additionally to this modular structure, all other operations are kept as vanilla as possible allowing users to easily change parts of the code (e.g., the splitting procedure).

Speed

As tree-based algorithms involve expensive loops over the training dataset, it is important that these computations are implemented in a compiled language. adaXT implements all computationally expensive operations in Cython. This results in speeds similar (although a few factors slower) than the corresponding scikit-learn implementations. However, due to its modular structure and the avoidance of technical speed-ups, adaXT does not intend to provide state-of-the-art speed and users mainly concerned with speed should consider more targeted implementations.

Intended Audience
- Science/Research
License
- OSI Approved :: BSD License
Operating System
- OS Independent
Programming Language
- Python :: 3

Download files

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

Source Distribution

adaxt-1.4.0.tar.gz (34.1 kB view details)

Uploaded Jan 17, 2025 Source

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

adaXT-1.4.0-cp313-cp313-win_amd64.whl (770.0 kB view details)

Uploaded Jan 17, 2025 CPython 3.13Windows x86-64

adaXT-1.4.0-cp313-cp313-win32.whl (644.8 kB view details)

Uploaded Jan 17, 2025 CPython 3.13Windows x86

adaXT-1.4.0-cp313-cp313-musllinux_1_2_x86_64.whl (5.3 MB view details)

Uploaded Jan 17, 2025 CPython 3.13musllinux: musl 1.2+ x86-64

adaXT-1.4.0-cp313-cp313-musllinux_1_2_i686.whl (5.0 MB view details)

Uploaded Jan 17, 2025 CPython 3.13musllinux: musl 1.2+ i686

adaXT-1.4.0-cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (5.2 MB view details)

Uploaded Jan 17, 2025 CPython 3.13manylinux: glibc 2.17+ x86-64

adaXT-1.4.0-cp313-cp313-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl (4.9 MB view details)

Uploaded Jan 17, 2025 CPython 3.13manylinux: glibc 2.17+ i686manylinux: glibc 2.5+ i686

adaXT-1.4.0-cp313-cp313-macosx_11_0_arm64.whl (820.8 kB view details)

Uploaded Jan 17, 2025 CPython 3.13macOS 11.0+ ARM64

adaXT-1.4.0-cp312-cp312-win_amd64.whl (772.8 kB view details)

Uploaded Jan 17, 2025 CPython 3.12Windows x86-64

adaXT-1.4.0-cp312-cp312-win32.whl (646.4 kB view details)

Uploaded Jan 17, 2025 CPython 3.12Windows x86

adaXT-1.4.0-cp312-cp312-musllinux_1_2_x86_64.whl (5.4 MB view details)

Uploaded Jan 17, 2025 CPython 3.12musllinux: musl 1.2+ x86-64

adaXT-1.4.0-cp312-cp312-musllinux_1_2_i686.whl (5.1 MB view details)

Uploaded Jan 17, 2025 CPython 3.12musllinux: musl 1.2+ i686

adaXT-1.4.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (5.2 MB view details)

Uploaded Jan 17, 2025 CPython 3.12manylinux: glibc 2.17+ x86-64

adaXT-1.4.0-cp312-cp312-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl (4.9 MB view details)

Uploaded Jan 17, 2025 CPython 3.12manylinux: glibc 2.17+ i686manylinux: glibc 2.5+ i686

adaXT-1.4.0-cp312-cp312-macosx_11_0_arm64.whl (833.6 kB view details)

Uploaded Jan 17, 2025 CPython 3.12macOS 11.0+ ARM64

adaXT-1.4.0-cp311-cp311-win_amd64.whl (775.2 kB view details)

Uploaded Jan 17, 2025 CPython 3.11Windows x86-64

adaXT-1.4.0-cp311-cp311-win32.whl (648.5 kB view details)

Uploaded Jan 17, 2025 CPython 3.11Windows x86

adaXT-1.4.0-cp311-cp311-musllinux_1_2_x86_64.whl (5.5 MB view details)

Uploaded Jan 17, 2025 CPython 3.11musllinux: musl 1.2+ x86-64

adaXT-1.4.0-cp311-cp311-musllinux_1_2_i686.whl (5.2 MB view details)

Uploaded Jan 17, 2025 CPython 3.11musllinux: musl 1.2+ i686

adaXT-1.4.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (5.4 MB view details)

Uploaded Jan 17, 2025 CPython 3.11manylinux: glibc 2.17+ x86-64

adaXT-1.4.0-cp311-cp311-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl (5.1 MB view details)

Uploaded Jan 17, 2025 CPython 3.11manylinux: glibc 2.17+ i686manylinux: glibc 2.5+ i686

adaXT-1.4.0-cp311-cp311-macosx_11_0_arm64.whl (826.2 kB view details)

Uploaded Jan 17, 2025 CPython 3.11macOS 11.0+ ARM64

adaXT-1.4.0-cp310-cp310-win_amd64.whl (774.4 kB view details)

Uploaded Jan 17, 2025 CPython 3.10Windows x86-64

adaXT-1.4.0-cp310-cp310-win32.whl (651.6 kB view details)

Uploaded Jan 17, 2025 CPython 3.10Windows x86

adaXT-1.4.0-cp310-cp310-musllinux_1_2_x86_64.whl (5.1 MB view details)

Uploaded Jan 17, 2025 CPython 3.10musllinux: musl 1.2+ x86-64

adaXT-1.4.0-cp310-cp310-musllinux_1_2_i686.whl (4.8 MB view details)

Uploaded Jan 17, 2025 CPython 3.10musllinux: musl 1.2+ i686

adaXT-1.4.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (4.9 MB view details)

Uploaded Jan 17, 2025 CPython 3.10manylinux: glibc 2.17+ x86-64

adaXT-1.4.0-cp310-cp310-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl (4.7 MB view details)

Uploaded Jan 17, 2025 CPython 3.10manylinux: glibc 2.17+ i686manylinux: glibc 2.5+ i686

adaXT-1.4.0-cp310-cp310-macosx_11_0_arm64.whl (826.6 kB view details)

Uploaded Jan 17, 2025 CPython 3.10macOS 11.0+ ARM64

File details

Details for the file adaxt-1.4.0.tar.gz.

File metadata

Download URL: adaxt-1.4.0.tar.gz
Upload date: Jan 17, 2025
Size: 34.1 kB
Tags: Source
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/6.0.1 CPython/3.12.8

File hashes

Hashes for adaxt-1.4.0.tar.gz
Algorithm	Hash digest
SHA256	`48732f773b4b66d0f73882156bd4efe7453ce6d329414e5f2f9b817f9d9989ef`
MD5	`66a77f14e7fea03dc6564095f3c50385`
BLAKE2b-256	`c9f7c8a79be7f3dce5538a1a33aa8380ab662ed9c0121015a03e91b59196ac77`

Algorithm	Hash digest
SHA256	`f808d7c17c7b8d67d2d91cf094d50f28b35b481853003f4e5956cfe4b2d64efd`
MD5	`3a291ffc8d612cf6d140edb3078d2bf0`
BLAKE2b-256	`abfff7e2b319e8fdb034ae3cc2ffb20c9277256258c8d1efc6c672744bf61364`

Algorithm	Hash digest
SHA256	`3567ddcb68f045280fcddf7be5e407de1538322b18fb5ddf49430ffff0041f48`
MD5	`c71615110b16143fc88c10841eaf5a57`
BLAKE2b-256	`8cb6ece13dcdecdd5fb311b567dd4d7795f3ffa722fa720d1b70fdc617f103f7`

Algorithm	Hash digest
SHA256	`e5e42d840c3ac357364b7f6be39adc892d7232f3559b49816262621132c1b28c`
MD5	`5b4d666730ddc5f3e7405d2d133976ad`
BLAKE2b-256	`8759a0fc79c929b6cc81e2a03a07af97bec0e1f2a857e8515445011070696980`

Algorithm	Hash digest
SHA256	`fc9b9004035e2b0b8f927e1fd2de79906f6030a376bc9bf69fa6b97d847d6de4`
MD5	`69b5823fca4aee9679c39aec32615e69`
BLAKE2b-256	`2ed4484dc1ee384b16c0a98a74066c139fe688a12343891c604e801b6d341907`

Algorithm	Hash digest
SHA256	`9d256fa03d4049ca3440751f9e6a29a29e599387a0a8d350731099110ea1cbc4`
MD5	`fc205a50e7f0fe39b9beff1ef467321e`
BLAKE2b-256	`c018ce7146a19e255b0f137689c50e289c297f1aabed33eb3cd7ba5233840955`

Algorithm	Hash digest
SHA256	`163fa8db302f514eb3afa2dea637e94385f1df5bbd08bfef2db4c564f14884cb`
MD5	`d5bb088fada55dfe7f7c1dffbe888e2e`
BLAKE2b-256	`ee8a753e5625b62b07cf968f01d7796bf08f7d11ec4dad2697e87fe92ef03534`

Algorithm	Hash digest
SHA256	`0f883eb4a7c3068e82220062759cc4d69378aa48792db0a5b605c9a3e7f5c491`
MD5	`bab2b12f9a686cef653fd3da6becf491`
BLAKE2b-256	`26a3dcf0f9e2733f8984ceb3253318d04b115c0248801a59d11c6c3a4c74c645`

Algorithm	Hash digest
SHA256	`11af329157943fbe39a2be88abde0bcb478d4ee2cee766bea441204d8577ff75`
MD5	`3d04b86799b35def4856c20af662c345`
BLAKE2b-256	`d42a8c9d415106b3da8069e9344e760b24191d82125491dcfa0dba30e9ea8072`

Algorithm	Hash digest
SHA256	`684120b7b3e968d6643d5e0d80ce78c25631c4a4e7558fd4cb7a39cfd3262e62`
MD5	`b5455eea87b0d93555c90203ebb04491`
BLAKE2b-256	`f1032eb75033b42f6d7826d2e1cf95471369d53188ac82e7f69b951dd2590a4b`

Algorithm	Hash digest
SHA256	`ccb21d52c6712fef89d4641877caa239f1833bf7f49976233de3a725c332d12a`
MD5	`5aefc3e91bd09f135194c395f4286457`
BLAKE2b-256	`c6df6f43106dcf1318bbfa51bd1a8a55921ba552bb80290fc929bdd8ef5ba45d`

Algorithm	Hash digest
SHA256	`375adfa93530ffad3c44ae186921e94df9deb812ef21dde26fc4f09497bc23ef`
MD5	`0c69d02c317d8d75b7872c72f8b0e908`
BLAKE2b-256	`9459f638bb6378a270c442c30179d09cc33218b8b6a22a4150abd593b80bf7af`

Algorithm	Hash digest
SHA256	`88cfeddd5852f188b9c7f56f1babff7f9c560ab929add7d06ce8330b917d5b22`
MD5	`5acd4c252351bb65e0cd18345b286d14`
BLAKE2b-256	`24772d60f36ea917c0ce4e18cdc321efbd525846b03d3428880efdf345a4d629`

Algorithm	Hash digest
SHA256	`8c7ca6faeb7a6884ebcfc65e29d24f07468e4f906725bcbe4b0899b304e402b4`
MD5	`21980cce54b490a7fffe23e0e0f6835e`
BLAKE2b-256	`74e8ffdaffb01f52a2447d88d14634d6961a8d2a46a2d827349451a59d525004`

Algorithm	Hash digest
SHA256	`9e59218a39c641b6d2ecf0d5ac93dad98dc0858288d4ec1e02b5909d157db97f`
MD5	`eebfb788423aca16712139428b505de5`
BLAKE2b-256	`920f614516f70ffaba849f562c2e985dead095e6ccb308823b8a298761ac3966`

Algorithm	Hash digest
SHA256	`dd889d1ab524d99a6a37fcaccf93f02b3a8f1c144e72c2bf18d12e1cdbd8ac01`
MD5	`f65e09ba95b8897cf8b30aa5cd101805`
BLAKE2b-256	`ab3f5df2d788552ea6b21806ea5c56698f8524324ab8fe7c8bc61a37ee45cb73`

Algorithm	Hash digest
SHA256	`c4cfc30efa84a37a9e9ef8485eb6e71a51fd60403d91dd46ebf0738e8fe196c5`
MD5	`f3481a394c9f76cd3e1397a4520ee88f`
BLAKE2b-256	`24a02c94b7087e50744766eb0610cacea6e5dd0a32aa5a03f8f102d598950433`

Algorithm	Hash digest
SHA256	`9bfa0d4483a4ee8e5ff1c265dcf66dc40da7995130d500c29160905545487611`
MD5	`86b250e8dfcafed42216ac5ff4baab46`
BLAKE2b-256	`660faf9106286641c278c2b26cd0f977b94846dc40511b9aa8870948bc2d8157`

Algorithm	Hash digest
SHA256	`f4314b4de3c1316da94f1418b58c16c68ba9200e766d054f0895d096b058844b`
MD5	`bd550b9ec0d2d228ecdbe99d5971f5cb`
BLAKE2b-256	`4e64c083b30764fe283740de2f6f7c75d38db71c29d8c6771363adf99cc3980f`

Algorithm	Hash digest
SHA256	`a956687abb85819744e1e8ef3b827fcefcc3f74c44034b6eb44d1a3bcda61238`
MD5	`236c8112e0bcd5002f20245d931f79fd`
BLAKE2b-256	`f7c90df2bd7ea39849cda1fd358cfbecc9ff2c7f654b905329ba3b970b2223c5`

Algorithm	Hash digest
SHA256	`5efa975f19174876554cec86d23a0f6145c24f4f8ed76581def3793a958f381a`
MD5	`eca5535e276c8cf08c0cc0832833c884`
BLAKE2b-256	`78bf312955e31908af7001b85f4fda8ad8d2a7a42eda1b1f7efd953351bd5dd0`

adaXT 1.4.0

Fast Adaptable and Extendable Trees for Research

Overview

Getting started

Project goals

Adaptable and extendable

Speed

Project links

Key dates

2 maintainers

GitHub Statistics

License

Requires

Provides Extra

Classifiers

Download files

Source Distribution

Built Distributions

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

Algorithm	Hash digest
SHA256	`0c74fec6eb7b4edc03b0fb1633954eada6d53b4369eecbdf00c71be9b3b2c840`
MD5	`305757361496794437788a88e9bcf752`
BLAKE2b-256	`6f87e606ac0503f29717e7c0934fff4a77dca0cce3425c9babe3d7079cf25063`

Algorithm	Hash digest
SHA256	`0c569636065d0b00d83a184cf810e0fac203a803f1d7f4288548ef5d3b737584`
MD5	`ea252c81a4de130a354b08457f9915e3`
BLAKE2b-256	`9cbd8f7cd10c1a1a88375fdd4b94810cca08d889638865ad4a8f19be56b209f4`

Algorithm	Hash digest
SHA256	`6b331d8f25a9b4ad6f40f5904ecfa6723fcf131d17535fc367112b79a719c9fb`
MD5	`f421a56d9c84c89dfd3e55033b97a5eb`
BLAKE2b-256	`0e71249577cebd28e2643124dc5af69c83f35fe616e66dec902ae2230dee7578`

Algorithm	Hash digest
SHA256	`bc88e160e81b128a8a0a6fbf0360fc7f75998ab9efac5578058c7301c61b2894`
MD5	`71b216f0e926d3a85e1910f0ea9d6aee`
BLAKE2b-256	`9f3a4f313019aa7904c56cf2d0207120c46d0043aed28c83bd009b8ac232d25a`

Algorithm	Hash digest
SHA256	`37567d1fd853c91b77d75047d2275903299f17282716058c7ade5fb081f02311`
MD5	`d57f69c2892be0b18420cca3d4a5b6c8`
BLAKE2b-256	`73db80d56cf253f254d439b5eeac9ed0cb012bf75b3388fb54cffa249230cc90`

Algorithm	Hash digest
SHA256	`57c17bd1f4ddd02d6861352ccc73bfc702d0dd60a8167b1b7c32b717b7700855`
MD5	`89738812b3d7fa65ba642e0608186730`
BLAKE2b-256	`27b9a3de0c0dd86a3680ee3fcf6d2776b7db33e9db3ecf2c06c30039e73d68c2`

Algorithm	Hash digest
SHA256	`a38b0e084df1526340255a2f870f3428979a2789ec8d377df7d3a3617c78b749`
MD5	`deee363ed4c5d5ff33b12c000ef089ae`
BLAKE2b-256	`fb75da4f2465d900236e4587fe9da074f6e7bb8f56ee033572bf851b72e679fa`

Algorithm	Hash digest
SHA256	`060e98e5c45249cf9d406c43f0edf1ec1e94cb8ea01b590cc85b4ecd26149066`
MD5	`35315042f8a36cfd180ade48e2fecede`
BLAKE2b-256	`a5b0a56f54b1d72c8f7813eb429728e0a8c2c86cea14299955e80577ffd8b234`