Prediction of ambient carbon monoxide concentration using nonlinear time series analysis technique

This study evaluates the potential of nonlinear time series analysis based methods in predicting the carbon monoxide concentration in an urban area. To establish the functional relationship between current and future observations, two models based on local approximations and neural network approximations are used. To compare the performance of the models, an autoregressive integrated moving average model is also applied. The multi-step forecasting capabilities of the models are evaluated.