Classy Ensemble: A Novel Ensemble Algorithm for Classification

TL;DR

Classy Ensemble is a new ensemble algorithm that improves classification accuracy by combining models based on per-class accuracy, outperforming existing methods across diverse datasets including images and large-scale data.

Contribution

Introduces Classy Ensemble, a novel weighted ensemble method, along with three enhancements integrating clustering, deep learning, and evolutionary algorithms for improved classification.

Findings

Outperforms existing ensemble methods on 153 datasets.

Effective on image datasets like CIFAR and ImageNet.

Enhanced with evolutionary algorithms for better performance on challenging data.

Abstract

We present Classy Ensemble, a novel ensemble-generation algorithm for classification tasks, which aggregates models through a weighted combination of per-class accuracy. Tested over 153 machine learning datasets we demonstrate that Classy Ensemble outperforms two other well-known aggregation algorithms -- order-based pruning and clustering-based pruning -- as well as the recently introduced lexigarden ensemble generator. We then present three enhancements: 1) Classy Cluster Ensemble, which combines Classy Ensemble and cluster-based pruning; 2) Deep Learning experiments, showing the merits of Classy Ensemble over four image datasets: Fashion MNIST, CIFAR10, CIFAR100, and ImageNet; and 3) Classy Evolutionary Ensemble, wherein an evolutionary algorithm is used to select the set of models which Classy Ensemble picks from. This latter, combining learning and evolution, resulted in improved performance on the hardest dataset.