Prediction of air pollutant concentration based on sparse response back-propagation training feedforward neural networks

摘要: In this paper, we predict air pollutant concentration using a feedforward artificial neural network inspired by the mechanism of human brain as useful alternative to traditional statistical modeling techniques. The is trained based on sparse response back-propagation in which only small number neurons respond specified stimulus simultaneously and provide high convergence rate for network, addition low energy consumption greater generalization. Our method evaluated Hong Kong monitoring station data corresponding meteorological variables five quality parameters were gathered at four stations over 4 years (2012–2015). results show that our training has more advantages terms precision prediction, effectiveness, generalization linear regression algorithms when compared with back-propagation.