Predicting atmospheric ozone using neural networks as compared to some statistical methods

Recent developments in the area of nonlinear nonparametric statistics show many similarities to artificial neural networks. Both fields share a common goal in data mining, in trying to extract information and patterns in data sets, and in fitting appropriate models to data for interpretation and prediction. In this paper, the authors present an empirical comparison of the predictive power on an atmospheric ozone data of several statistical models and neural networks. The statistical models are the classical linear regression model, general additive model-a generalised form of linear model, locally weighted regression model and projection pursuit regression model. The projection pursuit regression model resembles the neural net in many ways and it is interesting to find out that they both provide similar predictions