SO2 concentrations forecasting for different hours in advance for the city of Salamanca, Gto., Mexico.
This work presents a method to predict Sulphur Dioxide (SO2) concentrations for the city of Salamanca, Gto., based on real data provided by a Monitoring network during the months of December of years 2002, 2003 and 2004. A Generalized Regression Neural Network (GRNN) is used to perform the predictions. The GRNN was trained with the data obtained from years 2002 and 2003, and the data from year 2004 were the test group data. Predictions were made for 1, 12 and 24 hours ahead, and results have been promising. Root Mean Squared Error (RMSE) and Mean Absolute Error (MAE) are the performance measures for the GRNN.
This application is new for us and we are exploring new techniques to solve the problem of the estimate and the technique that you propose it would be interesting to follow it, we will take it in consideration and thank you for the feedback
Antonio Vega-Corona
Yes, as Dr. Vega-Corona says, we are new to the use of these techniques, and for some of the authors this was our first work. We haven't tried other Neural Network techniques to improve the Prediction performance either. Thanks for the question and for the clue about Fuzzy Logic. We will consider these techniques for our Further Work.
Regards,
Ulises S. Mendoza
Who's online
There are currently 0 users and 143 guests online.
A very interesting application of neural networks. I am wondering have you considered using nonlinear Neuro-Fuzzy prediction techniques? They have been applied in several domains from ground water level prediction to stock market prediction.