A learning algorithm using parallel neuron model.

This paper proposes a parallel neuron network model which integrates heterogeneous neural networks to conduct partial learning in each network. The learning algorithm is based on the LVQ (Linear Vector Quantization) algorithm of Kohonen (1990) for clustering, and ADALINE (Adaptive Linear Neuron) network of Widrow and Hoff (1960) for parallel learning. The proposed algorithm consists of two parts. First, N patterns to be learned are categorized into C clusters by LVQ clustering algorithm. Second, C patterns that were selected from each cluster of C are learned as input pattern of ADALINE. Data used in this paper was 1,124 samples of electronic signals generated from Inductive Loop Detectors (ILD). In the experiment, 807 out of 1124 vehicles were correctly recognized, showing 71.8% recognition ratio. This result is a 10.2% improvement over backpropagation algorithms.