Attention-based Random Forest and Contamination Model

TL;DR

This paper introduces ABRF, an attention-based random forest model that assigns trainable attention weights to trees based on instance similarity, improving regression and classification tasks.

Contribution

It proposes a novel ABRF model with three modifications, integrating attention mechanisms and optimization algorithms into random forests for enhanced performance.

Findings

ABRF outperforms traditional RF in experiments.

The models effectively incorporate attention weights based on instance proximity.

Modifications using optimization and gradient algorithms improve adaptability.

Abstract

A new approach called ABRF (the attention-based random forest) and its modifications for applying the attention mechanism to the random forest (RF) for regression and classification are proposed. The main idea behind the proposed ABRF models is to assign attention weights with trainable parameters to decision trees in a specific way. The weights depend on the distance between an instance, which falls into a corresponding leaf of a tree, and instances, which fall in the same leaf. This idea stems from representation of the Nadaraya-Watson kernel regression in the form of a RF. Three modifications of the general approach are proposed. The first one is based on applying the Huber's contamination model and on computing the attention weights by solving quadratic or linear optimization problems. The second and the third modifications use the gradient-based algorithms for computing trainable parameters. Numerical experiments with various regression and classification datasets illustrate the proposed method.