IJCRR - Vol 06 Issue 16, August, 2014
IMPROVED ARTIFICIAL NEURAL NETWORK PERFORMANCE ON SURFACE OZONE PREDICTION USING PRINCIPAL COMPONENT ANALYSIS
Author: K. Padma, R. Samuel Selvaraj, S. Arputharaj, B. Milton Boaz
Correlated sample data normally creates confusion over ANN (Artificial Neural network) during the learning process. In this work the Principal Component Analysis (PCA) method is used for elimination of correlated terms in data. After application of PCA, the uncorrelated input data were used to train a Multi-Layer Perceptron (MLP) ANN system. The results revealed that the elimination of correlated information by using the PCA method improved the ANN estimation performance. we measured the surface ozone and its influencing factors during the period June 2011- September 2012 at Chennai, a tropical site on the Southeast coast of India situated at 13 04’N 80 17’E. The input data that was used for building the network are the wind speed, temperature, relative humidity, UV radiation had been used in neural networks for the prediction of daily surface Ozone 24 hours in advance.
Keywords: Multi-layer perceptron, Artificial neural network, Principal component analysis, activation function, hidden layers, ‘PC variance’
K. Padma, R. Samuel Selvaraj, S. Arputharaj, B. Milton Boaz. IMPROVED ARTIFICIAL NEURAL NETWORK PERFORMANCE ON SURFACE OZONE PREDICTION USING PRINCIPAL COMPONENT ANALYSIS International Journal of Current Research and Review. Vol 06 Issue 16, August, 01-06
1. Akram Ali, Factors affecting on response of Broad bean and corn to air quality and soil CO2 flux rates in Egypt. Water Air soil Pollute.195, 311-323, 2008.
2. Andrew C. Comrie. Comparing neural network and regression models for Ozone forecasting. Air waste Management. Assoc. 47,653-663, 1997,.
3. Bandyopathyay.G , S.Chattopadhyay,. Single hidden artificial neural network models versus multiple linear regression model forecasting the time series ozone. Int. J. Envireon. Scxi. Tec., 4(1).,141-149,. 2007
4. Dovile Laurinaviene., Ground level Ozone Air pollution in Vilnius City. Environmental Researech, Engineering and Management, 2008, No. 3(49), 21-28.
5. Elkamel.A, S.Abdul-Wahab., W. Bouhamra, E. Alper., Measurement and prediction of Ozone levels around heaviloy industrialized area. A neural network approach.Adance in Environment Research 5, 47-59, 2000.
6. Elampari. K, Chithambarthanu T. Diurnal and seasonal variations in surface ozone levels at tropical Semi - Urban site, Nagercoil, India, and relationships with meteorological conditions. International Journal of Science and Technology, 2011, volume1, No.2.
7. Emberson. D., M.R. Ashmore, F. Murray, J.C.I. Kuylenstierna, K.E. Percy, T. Izuta, Y. Zheng, H. Shimizu, B.H. Sheu, C.P. Liu, M. Agrawal, A. Wahid, N.M. Abdel-Latif, M. van Tienhoven, L.I. de Bauer, M. Domingos, Impacts of air pollutants on vegetation in developing countries,Water Air Soil Pollut. 2001, 107–118.
8. Girish Kumar Jha ., Artificial Neural Networks., Indian Agriculturral Research Institute.Pusa , New Delhi
9. Joliffe. I.T, Principal component Analysis, Springer Verlag,1986, 533 – 536.
10. Junitha Mohamad-Saleh, Brian S. Hoyle., Improved Neural Networ performance using principal component analysis on matlab. International journal of the computer,vol.16. No.2. pp 1-8, 2008.
11. Lodhe A.L, D.B. Jadhav,P.S. Buchunde and M.J.Kartha. Surface Ozone variablility in the urban and nearby rural locations of tropical India. Currennt science, Volumw 95, No.12, 25. A13. 2008).
12. Michael Frei, Juan Pariasca Tanaka and Matthias Wissuwa. Genotypic variation in tolerance to elevated ozone in rice; dissection of distinct genetic factors linked to tolerance mechanisms, Journal of Experimental Botany, 2008, 13, 3741-3752.
13. Pulikesi M.,P. Baskaralingam, Ramamurthi.V., Sivanesan. Studies on surface ozone in Chennai. Research journal of chemistryand environment, 2005, Vol. 9., issue.
14. Samuel Selvaraj R., Milton Boaz, B., Sachithananthem C.P., Padma, K., Steephen Rajkumar S. Inbanathan., Kanmani RajaselviG. Indira and VImalpriya S.P., Measurement of surface ozone in the year 2011 at different sites over Tamil Nadu, India. Indian Journal of science and Technology, 2011, vol. 5, No. 2.
15. Samuel selvaraj R., Padma K, Miltoin Boaz B, Seasonal variation of surface ozone and its association with meteorological parameters, UV radiation, rainfall, cloud cover, over Chennai, India. Current science, vol. 105, no. 5, 10 september 2013.
16. Su Lee and shih-Wei Tsai, Passive sampling of ambient ozone by solid phase micro extraction with on – fiber derivatization, Analytica Chimica Acta, 2008, Vol.610, Issue 2, 149-155.
17. Varshney C.K., M. Aggarwal, Ozone pollution in the urban atmosphere of Delhi, Atmos. Environ. 1992, 26B, 3, 291– 294.
18. Wenjian Wang, Zongben Xu, Jane Weizhen Lu,. Three improved neural network models for air quality forecasting. Engineering computations, vol. 20, No. 2, 2003.