The Forecasting Of Water Level using Artificial neural Networks In Bekasi, West Java - Indonesia

Abstract

: Forecasting of a water level constitutes major input informationin flood forecasting, it would
provide a warning of the impending flood during the times of floods. Information on water levels is
collect e-data convenient stream gauging station by automatic water level recorders. T h i s s t u d y
a i m s t o a nalysis of these records and prediction of future events based on these records. Methods t
of or ecast water levels at a site a long a river are generally statistical model based. The Artificial
Neural Network (ANN) is applied for short-term prediction, especially for a variety of parameters such a s rainfall, land use, and river characteristics influence the water level in a highly non-linear manner. The results of this study shows the real-time forecasting of water levels at a chosen site continuously. The network is trained by using back propagation algorithms. The training results are verified with untrained data. The conclusion of this study are the suitability of an NN for flow prediction with high accuracy at a fraction of the computational time required by the water stage at gauge station models; and a technique for detecting less sensitive input neurons in an effort to reduce
unnecessary data collection