Ozone Modeling Using Neural Networks

摘要: Abstract Ozone models for the city of Tulsa were developed using neural network modeling techniques. The meteorological data from Oklahoma Mesonet and ozone, nitric oxide, nitrogen dioxide (NO2) Environmental Protection Agency monitoring sites in area. An initial model trained with only eight surface input variables NO2 was able to simulate ozone concentrations a correlation coefficient 0.77. then used evaluate sensitivity primary that affect concentrations. most important (NO2, temperature, solar radiation, relative humidity) showed response curves strong nonlinear codependencies. Incorporation previous 3 days into increased 0.82. As expected, correlated best recent (1-day previous) values. model’s i...