Model Penelusuran Banjir Menggunakan Pendekatan Adaptive Neuro Fuzzy Inference System (ANFIS) (Studi Kasus : Sub Das Siak)

Abstract

The purpose of this research is predict the height of water level of the Siak Sub-watershed Pantai Cermin Station in 2012 by using water level data recorded by Tandun Station and Pantai Cermin Station in the same year. The approach itself uses softcomputing method. Softcomputing method has been widely used as hydrological analysis model, one of them for flood routing forecasting. ANFIS model is one of the softcomputing method that can predict the height of water level. ANFIS models need to be tested reliability of the Siak sub-watershed given the importance of streamflow information to generate management, planning, and early warning system of flood. In this research, ANFIS model which built using AWLR water level data Station Tandun and Station Pantai Cermin for 4 years (2009-2012). ANFIS model was conducted by network model configuration is 2 input and 1 output. The result of water level forecasting by using ANFIS models show that the results of the training, testing and validation that excellent results with value the test statistic parameters of the correlation coefficient (R) is 0,9314 that is in category of very strong correlation, statistic parameters of RMSE (root mean square error) 0,4866 meter, and the test parameters of the average valuation relative error by 14,1439 %.
Keywords: Flood Routing, Water Level, Softcomputing, ANFIS.