Estimation of Background Ozone Temporal Profiles using Neural Networks

IEEE Press
Publication Type:
Conference Proceeding
Proceedings 2011 IEEE International Conference on Intelligent Computing and Intelligent Systems (ICIS 2011), 2011, pp. 292 - 297
Issue Date:
Filename Description Size
Thumbnail2010006278OK.pdf Published version898.83 kB
Adobe PDF
Full metadata record
It is recognised that effective determination of the background ozone level (BOL) is important to provide the reference level in which human health risk assessment can be undertaken. The concept of BOL may be easily understood but in practice it is hard to distinguish between natural and anthropogenic effects. Apart from existing approaches to the BOL determination, a new quantisation method will be presented in this work, by evaluating the ozone versus nitric oxide (O3-NO) relationship to estimate the BOL mean value. To this end, a computational intelligence approach using the radial basis function neural network (RBFNN) is proposed for temporal estimation of the background ozone level. An improved method called forward selection with weighted least squares (FSWLS) will be introduced to select the network centres. This can beneficially result in a minimal number of hidden neurons used, especially when dealing with noisy data. The developed neural network will be utilised to map the non-linear relationship between ozone precursors and other factors in ozone generation as the inputs, with the background ozone level as the output. The resulting metamodel, subject to some statistical criteria, demonstrates its capability of estimating the background ozone temporal profiles with a reasonably good accuracy.
Please use this identifier to cite or link to this item: