Greedy Function Approximation: A Gradient Boosting Machine

DMCA

Greedy Function Approximation: A Gradient Boosting Machine (2000)

Cached

Download Links

Other Repositories/Bibliography

CiteULike

by Jerome H. Friedman

Venue:	Annals of Statistics
Citations:	997 - 13 self

BibTeX

@ARTICLE{Friedman00greedyfunction,
    author = {Jerome H. Friedman},
    title = {Greedy Function Approximation: A Gradient Boosting Machine},
    journal = {Annals of Statistics},
    year = {2000},
    volume = {29},
    pages = {1189--1232}
}

OpenURL

Abstract

Function approximation is viewed from the perspective of numerical optimization in function space, rather than parameter space. A connection is made between stagewise additive expansions and steepest{descent minimization. A general gradient{descent \boosting" paradigm is developed for additive expansions based on any tting criterion. Specic algorithms are presented for least{squares, least{absolute{deviation, and Huber{M loss functions for regression, and multi{class logistic likelihood for classication. Special enhancements are derived for the particular case where the individual additive components are regression trees, and tools for interpreting such \TreeBoost" models are presented. Gradient boosting of regression trees produces competitive, highly robust, interpretable procedures for both regression and classication, especially appropriate for mining less than clean data. Connections between this approach and the boosting methods of Freund and Shapire 1996, and Frie...

Keyphrases

greedy function approximation gradient boosting machine regression tree function approximation function space gradient boosting general gradient descent numerical optimization particular case special enhancement clean data descent minimization tting criterion huber loss function absolute deviation additive expansion specic algorithm multi class logistic likelihood treeboost model interpretable procedure parameter space stagewise additive expansion individual additive component

DMCA