Abstract—The management of water resources is important to prevent water problems: floods and water shortages. The foreknowledge allows time for officials to sufficient preparation to deal with the problem. This study aims to determine the appropriate weight for predicting runoff from the merge of runoff prediction results from two algorithms: Artificial Neural Network and Support Vector Regression with linear regression modeling. In this paper, we compare the runoff predictive performance of the three algorithms: Linear Regression, Artificial Neural Network, and Support Vector Regression. We use remote sensing data, which are the Normalized Difference Vegetation Index (NDVI) obtained from the NOAA STAR. The ground station rainfall, runoff, the number of rainy days and temperature data in Mun basin, Thailand, are obtained from the Meteorological Department. We evaluate the model performance using two statistical values: Correlation Coefficient and Root Mean Squared Error. Experimental results confirm the best performance of our proposed method.
Index Terms—Runoff, artificial neural network, support vector regression.
R Chanklan is with the School of Computer Engineering, Suranaree University of Technology (SUT), 111 University Avenue, Muang, Nakhon Ratchasima 30000, Thailand (e-mail: arc_angle@hotmail.com).
K. Chaiyakhan is with the Computer Engineering Department, Rajamangala University of Technology Isan, Nakhon Ratchasima, Thailand (e-mail: kedkarnc@hotmail.com).
K. Kerdprasop is with the School of Computer Engineering and the Head of Knowledge Engineering Research Unit, SUT, Thailand (e-mail: kerdpras@sut.ac.th).
N. Kerdprasop is with the School of Computer Engineering and the Head of Data Engineering Research Unit, SUT, Thailand (e-mail: nittaya@sut.ac.th).
[PDF]
Cite: Ratiporn Chanklan, Kedkarn Chaiyakhan, Kittisak Kerdprasop, and Nittaya Kerdprasop, "A Hybrid Modeling Technique to Predict Runoff," International Journal of Modeling and Optimization vol. 7, no. 2, pp. 60-64, 2017.
