• Corpus ID: 26037613

CatBoost: gradient boosting with categorical features support

  title={CatBoost: gradient boosting with categorical features support},
  author={Anna Veronika Dorogush and Vasily Ershov and Andrey Gulin},
In this paper we present CatBoost, a new open-sourced gradient boosting library that successfully handles categorical features and outperforms existing publicly available implementations of gradient boosting in terms of quality on a set of popular publicly available datasets. The library has a GPU implementation of learning algorithm and a CPU implementation of scoring algorithm, which are significantly faster than other gradient boosting libraries on ensembles of similar sizes. 

Figures and Tables from this paper

Factorized MultiClass Boosting

A new approach to multiclass classification problem that decomposes the problem into a series of regression tasks, that are solved with CART trees, allowing to reach high-quality results in significantly less time without class re-balancing.

MP-Boost: Minipatch Boosting via Adaptive Feature and Observation Sampling

Boosting methods are among the best generalpurpose and off-the-shelf machine learning approaches, gaining widespread popularity. In this paper, we seek to develop a boosting method that yields

Wide Boosting

This paper presents a simple adjustment to GB that allows the output of a GB model to have increased dimension prior to being fed into the loss and is thus "wider" than standard GB implementations.

StructureBoost: Efficient Gradient Boosting for Structured Categorical Variables

The resulting package, called StructureBoost, is shown to outperform established packages such as CatBoost and LightGBM on problems with categorical predictors that contain sophisticated structure and can make accurate predictions on unseen categorical values due to its knowledge of the underlying structure.

Multi-Target XGBoostLSS Regression

An exten- sion of XGBoostLSS is presented that models multiple targets and their dependencies in a probabilistic regression setting that outperforms existing GBMs with respect to runtime and compares well in terms of accuracy.

Competitive Analysis of the Top Gradient Boosting Machine Learning Algorithms

This research performs an exhaustive 360 degree comparative analysis of each of the four state-of-the-art gradient boosting algorithms viz.

Challenges and Opportunities of Building Fast GBDT Systems

This survey paper reviews the recent GBDT systems with respect to accelerations with emerging hardware as well as cluster computing, and compares the advantages and disadvantages of the existing implementations.

Gradient Boosting Machine with Partially Randomized Decision Trees

This work proposes to apply the partially randomized trees which can be regarded as a special case of the extremely randomized trees applied to the gradient boosting machine to reduce the computational complexity of the gradient boost machine.

Hyperboost: Hyperparameter Optimization by Gradient Boosting surrogate models

This paper proposes a new surrogate model based on gradient boosting, where it uses quantile regression to provide optimistic estimates of the performance of an unobserved hyperparameter setting, and combines this with a distance metric between unobserved and observedhyperparameter settings to help regulate exploration.

agtboost: Adaptive and Automatic Gradient Tree Boosting Computations

agtboost is an R package implementing fast gradient tree boosting computations in a manner similar to other established frameworks such as xgboost and LightGBM, but with significant decreases in



CatBoost: unbiased boosting with categorical features

This paper presents the key algorithmic techniques behind CatBoost, a new gradient boosting toolkit and provides a detailed analysis of this problem and demonstrates that proposed algorithms solve it effectively, leading to excellent empirical results.

Fighting biases with dynamic boosting

Experimental results demonstrate that the open-source implementation of gradient boosting that incorporates the proposed algorithm produces state-ofthe-art results outperforming popular gradient boosting implementations.

XGBoost: A Scalable Tree Boosting System

This paper proposes a novel sparsity-aware algorithm for sparse data and weighted quantile sketch for approximate tree learning and provides insights on cache access patterns, data compression and sharding to build a scalable tree boosting system called XGBoost.

An empirical comparison of supervised learning algorithms

A large-scale empirical comparison between ten supervised learning methods: SVMs, neural nets, logistic regression, naive bayes, memory-based learning, random forests, decision trees, bagged trees, boosted trees, and boosted stumps is presented.

Winning The Transfer Learning Track of Yahoo!'s Learning To Rank Challenge with YetiRank

A novel pairwise method called YetiRank is introduced that modifies Friedman's gradient boosting method in part of gradient computation for optimization and takes uncertainty in human judgements into account and allowed yetiRank to outperform many state-of-the-art learning to rank methods in offline experiments.

Adapting boosting for information retrieval measures

This work presents a new ranking algorithm that combines the strengths of two previous methods: boosted tree classification, and LambdaRank, and shows how to find the optimal linear combination for any two rankers, and uses this method to solve the line search problem exactly during boosting.

GPU-acceleration for Large-scale Tree Boosting

A novel massively parallel algorithm for accelerating the decision tree building procedure on GPUs, which is a crucial step in Gradient Boosted Decision Tree (GBDT) and random forests training and can be used as a drop-in replacement for histogram construction in popular tree boosting systems to improve their scalability.

Stochastic gradient boosting

Enhancing LambdaMART Using Oblivious Trees

Experimental results suggest that the performance of the current state-of-the-art learning to rank algorithm LambdaMART can be improved if standard regression trees are replaced by oblivious trees and demonstrate that the use of oblivious trees can improve the performance by more than $2.2\%$.

Greedy function approximation: A gradient boosting machine.

A general gradient descent boosting paradigm is developed for additive expansions based on any fitting criterion, and specific algorithms are presented for least-squares, least absolute deviation, and Huber-M loss functions for regression, and multiclass logistic likelihood for classification.