A K-means-based Multi-subpopulation Particle Swarm Optimization for Neural Network Ensemble

TL;DR

This paper introduces KMPSO, a novel multi-subpopulation particle swarm optimization method that uses k-means clustering to train neural network ensembles, balancing diversity and accuracy effectively.

Contribution

It proposes a dynamic clustering approach within PSO to improve neural network ensemble training, which is a new integration of k-means with PSO for this purpose.

Findings

Achieves competitive performance on benchmark problems.

Effectively balances diversity and accuracy in ensembles.

Demonstrates improved training efficiency.

Abstract

This paper presents a k-means-based multi-subpopulation particle swarm optimization, denoted as KMPSO, for training the neural network ensemble. In the proposed KMPSO, particles are dynamically partitioned into clusters via the k-means clustering algorithm at every iteration, and each of the resulting clusters is responsible for training a component neural network. The performance of the KMPSO has been evaluated on several benchmark problems. Our results show that the proposed method can effectively control the trade-off between the diversity and accuracy in the ensemble, thus achieving competitive results in comparison with related algorithms.