Project details for XGBoost

Logo XGBoost v0.1

by crowwork - March 27, 2014, 07:09:52 CET [ Project Homepage BibTeX Download ]

view (9 today), download ( 1 today ), 3 subscriptions


xgboost: eXtreme Gradient Boosting

This is yet another gradient boosting (tree) (GBRT) library.

The implementation is based on C++, optimized for memory and efficiency, to try target large data with a single machine.

Palatalization is done with OpenMP. The algorithm relies on sparse feature format, which means it naturally handles missing values.

Supported key components so far:

  • Gradient boosting models:

    • regression tree (GBRT)

    • linear model/lasso

  • Objectives to support tasks:

    • regression

    • classification

  • OpenMP implementation

Changes to previous version:

Initial Announcement on

BibTeX Entry: Download
URL: Project Homepage
Supported Operating Systems: Linux, Mac Os X
Data Formats: Libsvm
Tags: Parallel, Gradient Boosting, Tree, Ensemble Learning
Archive: download here

Other available revisons

Version Changelog Date
  • Distributed version of xgboost that runs on YARN, scales to billions of examples

  • Direct save/load data and model from/to S3 and HDFS

  • Feature importance visualization in R module, by Michael Benesty

  • Predict leaf index

  • Poisson regression for counts data

  • Early stopping option in training

  • Native save load support in R and python

  • xgboost models now can be saved using save/load in R

  • xgboost python model is now pickable

  • sklearn wrapper is supported in python module

  • Experimental External memory version

May 12, 2015, 08:57:16

New features in the lastest changes

  • Distributed version now runs on Hadoop YARN
March 9, 2015, 23:17:29

New features in the lastest changes

  • Distributed version that scale xgboost to even larger problems with cluster

  • Feature importance visualization in R module

  • Predict leaf index

January 21, 2015, 19:33:24

New features: - R support that is now on CRAN

  • Faster tree construction module

  • Support for boosting from initial predictions

  • Linear booster is now parallelized, using parallel coordinated descent.

September 2, 2014, 02:43:31

New features: - Python interface - New objectives: weighted training, pairwise rank, multiclass softmax - Comes with example script on Kaggle Higgs competition, 20 times faster than skilearn's GBRT

May 17, 2014, 07:27:59

Initial Announcement on

March 27, 2014, 07:09:52


No one has posted any comments yet. Perhaps you'd like to be the first?

Leave a comment

You must be logged in to post comments.