Abstract:Random Forests have been one of the most popular bagging methods in the past few decades, especially due to their success at handling tabular datasets. They have been extensively studied and compared to boosting models, like XGBoost, which are generally considered more performant. Random Forests adopt several simplistic assumptions, such that all samples and all trees that form the forest are equally important for building the final model. We introduce Enhanced Random Forests, an extension of vanilla Random Forests with extra functionalities and adaptive sample and model weighting. We develop an iterative algorithm for adapting the training sample weights, by favoring the hardest examples, and an approach for finding personalized tree weighting schemes for each new sample. Our method significantly improves upon regular Random Forests across 15 different binary classification datasets and considerably outperforms other tree methods, including XGBoost, when run with default hyperparameters, which indicates the robustness of our approach across datasets, without the need for extensive hyperparameter tuning. Our tree-weighting methodology results in enhanced or comparable performance to the uniformly weighted ensemble, and is, more importantly, leveraged to define importance scores for trees based on their contributions to classifying each new sample. This enables us to only focus on a small number of trees as the main models that define the outcome of a new sample and, thus, to partially recover interpretability, which is critically missing from both bagging and boosting methods. In binary classification problems, the proposed extensions and the corresponding results suggest the equivalence of bagging and boosting methods in performance, and the edge of bagging in interpretability by leveraging a few learners of the ensemble, which is not an option in the less explainable boosting methods.

RandomBoost: Simplified Multiclass Boosting Through Randomization

Boosting, Voting Classifiers and Randomized Sample Compression Schemes

A Scalable Stagewise Approach to Large-Margin Multiclass Loss-Based Boosting

A scalable direct formulation for multi-class boosting

A Direct Approach to Multi-class Boosting and Extensions

A scalable stage-wise approach to large-margin multi-class loss based boosting

Random-projection ensemble classification

Random projections: data perturbation for classification problems

Random Projections for Classification: A Recovery Approach

Multiclass Boosting: Simple and Intuitive Weak Learning Criteria

Binary Classification: Is Boosting stronger than Bagging?

The Many Faces of Optimal Weak-to-Strong Learning

Online Multiclass Boosting

Approximation and generalization properties of the random projection classification method

Boosting Probabilistic Reasoning Model for Face Recognition

Totally Corrective Multiclass Boosting with Binary Weak Learners

Fair MP-BOOST: Fair and Interpretable Minipatch Boosting

Experiments with a New Boosting Algorithm

A boosting method to select the random effects in linear mixed models

A Method to Boost Naïve Bayesian Classifiers

Random boosting and random^2 forests -- A random tree depth injection approach