An artificial neural network system with a backpropagation method can be used to predict next year's precipitation rainfall based on studies in previous three-year precipitation data by using output errors to calibrate the output to get the desired result. Research is using backpropagation methods to predict rainfall in the city of Ambon. The training data shows that the 2020 training data forecasts using 2019 data equals the 2020 precipitation data because it can be said that the training data has 100% accuracy. Using a system on the training data done data testing for 2021 predictions by using 2020 data it is close to 2021 precipitation data with an accuracy of 80% with an MSE = 0.070945. Thus, a 2022 precipitation prediction can be expected to have 90% accuracy from the average accuracy of training data and testing data.
Copyrights © 2022