Prediction water level is still done using statistical methods. However, problems will arise if the research done on dynamic systems, as well as the water level prediction system. Artificial neural network technology can properly identify data patterns of a dynamic system. This study was conducted to predict the height of water in Bogor using rainfall data, evaporation and water level, observations in 2009-2010. The design of the prediction model using neural networks backpropagation with MATLAB software. Characteristics of artificial neural networks used are: 1 input layer with two neurons (precipitation and evaporation), one hidden layer and one output layer (water level), the value of the learning rate of 0.9; momentum of 0.1; 3 hidden neurons, error tolerance of 0.0001 and a maximum of 10000 epoch The experimental results by looping earned 10 times the average error was 10.93%, and it can be concluded that the system can properly predict the height of the water
Copyrights © 2017