On using Adaptive Hybrid Intelligent Systems in PM10 Prediction
S. A. Asklany1, Khaled Elhelow2, M .Abd El-Wahab3

1Dr. S. A. Asklany, Department of Computer Science, Modern Academy Maadi, Cairo, Egypt.
2Dr. Khaled Elhelow, Department of Mathematics, Elmagmaa University, Elmagmaa, Saudi Arabia.
3Prof. M.Abd El-Wahab, Astronomy and Meteorology department, Cairo University, Cairo, Egypt.

Manuscript received on August 05, 2016. | Revised Manuscript received on August 06, 2016. | Manuscript published on September 05, 2016. | PP: 54-59 | Volume-6 Issue-4, September 2016. | Retrieval Number: D2905096416/2016©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: A comparative study based on producing two intelligence systems applied to PM10 prediction was presented in this work. Adaptive Network –based Fuzzy Inference System (ANFIS) used in build a system has three weather elements as input variables (Wind Speed, Wind Direction and Temperature) and the PM10 as output variable for PM10 nowcast model. Another technique used ANFIS in prediction of chaotic time series to get 6 hours forecast for PM10 from the present data. For developing the models, thirteen years hourly data for Mansoria station coordinates 29° 300′ 0″ N, 45° 45′ 0″ E from 1995to 2007 has been used. Different models employing a different training and testing data sets had been studied. The criteria of performance evaluation are calculated for estimating and comparing the performances of the two techniques. The results show that the two presented models success tools in PM10 prediction with acceptable root mean square error (RMSE); the model built on using ANFIS for chaotic time series prediction has smaller error compared with the adaptive network fuzzy inference system.
Keywords: Air quality, artificial intelligence, pollution, ANFIS, soft computing