Staged Neural Modeling with Application to Prediction of NOx Pollutant Concentrations in Urban Air

Abstract

AbstractAddressing the drawbacks of widely used forward neural network growing methods in neural modeling of time series and nonlinear dynamic systems, a staged algorithm is proposed in this paper for modeling and prediction of NOx Pollutant Concentrations in urban air in Belfast, Northern Ireland, using generalized single-layer network. In this algorithm, forward method is used for neural network growing, the resultant network is then refined at the second stage to remove inefficient hidden nodes. Application study confirms the effectiveness of the proposed method.KeywordsNitric OxideNeural ModelingNeural Network PredictionNonlinear System IdentificationCandidate PoolThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.