Improving generalization of radial basis function network with adaptive multi-objective particle swarm optimization

Abstract

In this paper, an adaptive evolutionary multiobjective selection method of RBF Networks structure is discussed. The candidates of RBF Network structures are encoded into particles in Particle Swarm Optimization (PSO). These particles evolve toward Pareto-optimal front defined by several objective functions with model accuracy and complexity. The problem of unsupervised and supervised learning is discussed with Adaptive Multi-Objective PSO (AMOPSO). This study suggests an approach of RBF Network training through simultaneous optimization of architectures and weights withAdaptive PSO-based multi-objective algorithm. Our goal is to determine whether Adaptive Multi-objective PSO can train RBF Networks, and the performance is validated on accuracy and complexity. The experiments are conducted on two benchmark datasets obtained from the machine learning repository. The results show that our proposed method provides an effective means for training RBF Networks that is competitive with PSObased multi-objective algorithm.

On a Multiobjective Training Algorithm for RBF Networks Using Particle Swarm Optimization

“…In this study, the construction of the Pareto RBF Network set obtained from the perspective of the MOBJ optimization is considered. The same approach was proposed in [21]. However, there are some important differences between the approach proposed in this work and the one proposed in [21].…”

“…The same approach was proposed in [21]. However, there are some important differences between the approach proposed in this work and the one proposed in [21]. First, this paper perceives the complexity defined as a Q-norm, which considers simultaneously the linear and nonlinear terms as components of the complexity.…”

A Novel Incremental Instruct Dynamic Intrusion Detection System Using PSO-RBF

“…In PSO, physical position is not an important factor. The population (or swarm) and the member particles are initialized by assigning random positions, velocities, and potential solutions are then "flown" through the hyperspace [7]- [10].…”