PARALLEL DISTRIBUTED NEURAL NETWORKS FOR CLASSIFICATION

Abstract

Neural networks have been parallelised in many different ways, but most of these methods involve the parallelisation of the internal looping operations of the models, maintaining a single neural network solution. This paper introduces a new method which involves solving a single classification problem with multiple neural networks, as such, the solution is derived by concurrently operating neural networks. The conglomeration of neural networks function together to provide a single classification solution. A generic waveform experiment is used to illustrate the effectiveness of the Parallel Distributed Neural Networks (PDNN) paradigm.

KEY WORDS:

• Parallel Distributed Neural Networks (PDNN)
• East Learning Artificial Neural Network (FLANN)
• Multilayer Perception (MLP)