Understanding and forecasting atmospheric quality parameters with the aid of ANNs

A problem solving domain for the application of artificial intelligence (AI) methods towards knowledge discovery for the purposes of modelling and forecasting is urban air quality. This domain has the specific characteristic that the key parameters of interest (pollutant concentration criteria) have multiple temporal (and spatial) scales. The present paper applies ANNs for the operational forecasting of the 8-hour running average for Ozone, 24 hours in advance, for two locations in Athens, Greece. Results verify the ability of the methods to analyze and model this knowledge domain and to forecast the levels of key parameters that provide direct input to the environmental decision making process.