https://github.com/mainro/xgbtune

a library to tune xgboost models
https://github.com/mainro/xgbtune

automl automl-algorithms gradient-boosting hyperparameter-optimization hyperparameter-tuning machine-learning parameter-tuning tuning-parameters xgboost

Last synced: 8 days ago
JSON representation

a library to tune xgboost models

• Host: GitHub
• URL: https://github.com/mainro/xgbtune
• Owner: MainRo
• Created: 2020-01-30T22:41:19.000Z (almost 5 years ago)
• Default Branch: master
• Last Pushed: 2020-02-14T14:47:28.000Z (over 4 years ago)
• Last Synced: 2024-10-31T06:51:38.882Z (15 days ago)
• Topics: automl, automl-algorithms, gradient-boosting, hyperparameter-optimization, hyperparameter-tuning, machine-learning, parameter-tuning, tuning-parameters, xgboost
• Language: Python
• Size: 74.2 KB
• Stars: 8
• Watchers: 3
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.rst

Awesome Lists containing this project

README

        
==========

XGBTune

==========

.. image:: https://badge.fury.io/py/xgbtune.svg

    :target: https://badge.fury.io/py/xgbtune

.. image:: https://github.com/mainro/xgbtune/workflows/Python%20package/badge.svg

    :target: https://github.com/mainro/xgbtune/actions?query=workflow%3A%22Python+package%22

    :alt: Github WorkFlows

.. image:: https://readthedocs.org/projects/xgbtune/badge/?version=latest

    :target: https://xgbtune.readthedocs.io/en/latest/?badge=latest

    :alt: Documentation Status

XGBTune is a library for automated XGBoost model tuning. Tuning an XGBoost

model is as simple as a single function call.

Get Started

============

.. code:: python

    from xgbtune import tune_xgb_model

    params, round_count = tune_xgb_model(params, x_train, y_train)

Install

========

XGBTune is available on PyPi and can be installed with pip:

.. code:: console

    pip install xgbtune

Tuning steps

=============

The tuning is done in the following steps:

* compute best round

* tune max_depth and min_child_weight

* tune gamma

* re-compute best round

* tune subsample and colsample_bytree

* fine tune subsample and colsample_bytree

* tune alpha and lambda

* tune seed

This steps can be repeated several times. By default, two passes are done.