Flood Detection Using Backpropagation Neural Network Method

Lack of river and watershed management will cause problems disasters. One it is the flood that can physical, social economic loss. So countermeasures or anticipation are needed by using Early Warning System (EWS) to provide early information if a going occur. This study uses five input indicators: temperature, humidity, water discharge, surface altitude rainfall data produce output in form notifications alarms for (EWS). Then configuration be processed Backpropagation Neural Network. Data used recorded real-time on research object two weeks with composition training testing percentage 80% 20%. The best backpropagation neural network model has 5 neurons layer architecture, 15 as hidden three layer. prediction result Network method, an RMSE score performance 2.16e-21 success system 91.33%. It shows excellent accuracy level.