MSI based machine learning algorithms
Project description
msitrees
msitrees is a set of machine learning models based on minimum surfeit and inaccuracy decision tree algorithm. So whats cool about them? No hyperparameters to optimize for base learner. Tree is regularized internally to avoid overfitting by design. Quoting authors of the paper:
To achieve this, the algorithm must automatically understand when growing the decision tree adds needless complexity, and must measure such complexity in a way that is commensurate to some prediction quality aspect, e.g., inaccuracy. We argue that a natural way to achieve the above objectives is to define both the inaccuracy and the complexity using the concept of Kolmogorov complexity.
For convenience, msitrees comes with scikit-learn style API and can be used with sklearn functions accepting estimator
object as parameter.
Installation
With pip
pip install msitrees
From source
git clone https://github.com/xadrianzetx/msitrees.git
cd msitrees
python setup.py install
Windows builds require at least MSVC2015
Quick start
from msitrees.tree import MSIDecisionTreeClassifier
from sklearn.datasets import load_iris
from sklearn.model_selection import cross_val_score
data = load_iris()
clf = MSIDecisionTreeClassifier()
cross_val_score(clf, data['data'], data['target'], cv=10)
# array([1. , 1. , 1. , 0.93333333, 0.93333333,
# 0.8 , 0.93333333, 0.86666667, 0.8 , 1. ])
Reference documentation
API documentation is available on here.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distributions
Built Distributions
Hashes for msitrees-0.1-cp38-cp38-win_amd64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | d34c2548ad69f2e72a7a2010345b3530ba172fb601dc468efc83e03327fbb0ed |
|
MD5 | ad4e511941c14e9b3e8ba73f1495dde4 |
|
BLAKE2b-256 | 722c2b91410b4c59af871a9e1c4e49d6932f16cee8ce054a484dcf7fa07fea91 |
Hashes for msitrees-0.1-cp37-cp37m-win_amd64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | c11409f147d9b8e522b9bef842b426784b7f036cb0512dbeb513e437d9588045 |
|
MD5 | aa2fc42bca2f2de2f510114296027759 |
|
BLAKE2b-256 | 25fcee276d42b00e71ca852a143c45a3cda63c07fab49f97292fd73ff10ca2bd |
Hashes for msitrees-0.1-cp37-cp37m-manylinux2010_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | d85307d6a2333df7ac239c98f90c00ff75cafba57219207d419500874e174ea6 |
|
MD5 | 8082bd15bf0aeea3f9d088bc46beba5b |
|
BLAKE2b-256 | 916fad3b5e5946bbfcd0598514641eb43c394a9a363144c8884541b57665879d |
Hashes for msitrees-0.1-cp36-cp36m-win_amd64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 96b044765cf64bcea97bc945872162e794dbcfc87d73cb0495d7053446270706 |
|
MD5 | 230c0e33966ce83f96ee3684cbe79604 |
|
BLAKE2b-256 | edf08b086c280382582e7ec5b5e78336bd57df9c7e1ee012c43fabe765e14597 |
Hashes for msitrees-0.1-cp36-cp36m-manylinux2010_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 68b6c9ca20bcb75124834c08beaf0a1c6060851ef7bb24eacfad738f40e105c6 |
|
MD5 | 0aa3d7e75fbd99b1d18f1ea10387559e |
|
BLAKE2b-256 | 2edd73342c606e614af462ecf4698d794a9153bf3e1840f5dcf89285c74dd6a8 |
Hashes for msitrees-0.1-cp35-cp35m-win_amd64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 8750f9182c8e389d80509c014ee5ff74ff2ed412a1796a4b852bb7abbfcf5c38 |
|
MD5 | 6c8e8f86737d0c60228d51a6a4ce31a7 |
|
BLAKE2b-256 | d523dc515b297c5cce6da01dd474b7fead9e5f1df5ca39006e24ebc5681ff080 |
Hashes for msitrees-0.1-cp35-cp35m-manylinux2010_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 67726ebf191b360aeaa49bb808cbe260093adbb139110ba82883d017af13c0c8 |
|
MD5 | 06e0a598b4fb7bef51168cf9e95f5755 |
|
BLAKE2b-256 | dc30182cd815f9529081c4ad7ca07338cc6ccbfab40cca1d02ca7ae174577abd |