Crossbreeding in Random Forest

TL;DR

This paper introduces Crossbred Random Forest (CRF), a novel method that enhances RF's speed and space efficiency by crossbreeding the best tree branches, maintaining accuracy while reducing model size.

Contribution

The paper proposes a new crossbreeding technique for RF that improves efficiency without sacrificing classification performance.

Findings

CRF outperforms standard RF in accuracy on tested datasets.

CRF reduces the number of trees needed for similar performance.

CRF maintains classification measures while improving speed and space efficiency.

Abstract

Ensemble learning methods are designed to benefit from multiple learning algorithms for better predictive performance. The tradeoff of this improved performance is slower speed and larger size of ensemble learning systems compared to single learning systems. In this paper, we present a novel approach to deal with this problem in Random Forest (RF) as one of the most powerful ensemble methods. The method is based on crossbreeding of the best tree branches to increase the performance of RF in space and speed while keeping the performance in the classification measures. The proposed approach has been tested on a group of synthetic and real datasets and compared to the standard RF approach. Several evaluations have been conducted to determine the effects of the Crossbred RF (CRF) on the accuracy and the number of trees in a forest. The results show better performance of CRF compared to RF.