Niche Particle Swarm Optimization for Neural Network Ensembles

This research investigates a swarm intelligence based multiobjective optimization algorithm for optimizing the behavior of a group of Arti cial Neural Networks (ANNs), where each ANN specializes to solving a speci c part of a task, such that the group as a whole achieves an e ective solution. Niche Particle Swarm Optimization (NichePSO) is a speciation technique that has proven e ective at locating multiple solutions in complex multivariate tasks. This research evaluates the e cacy of the NichePSO method for training a group of ANNs that form a neural network ensemble (NNE) for the purpose of solving a set of multivariate tasks. NichePSO is compared with a gradient descent method for training a set of individual ANNs to solve di erent parts of a multivariate task, and then combining the outputs of each ANN into a single solution. To date, there has been little research that has compared the e ectiveness of applying NichePSO versus more traditional supervised learning methods for the training of neural network ensembles.