Share
Export Citation
Flood Early Warning System Using River Water Level Prediction with Artificial Neural Network (Case Study: Jakarta City)
Saputro D.B.
Proceeding IEEE 8th Information Technology International Seminar Itis 2022
Abstract
The problem of flooding is still an important topic to be solved in Indonesia. In 2021, floods were the most common disaster. So, this research proposes one way to deal with flood disasters: by predicting the arrival of the flood by predicting the river’s water level to know the alert status in each river so that residents can anticipate the arrival of floods. In this study, the prediction of river water height will use one of the machine learning algorithms, Artificial Neural Network (ANN). The machine learning model that has been created will be presented using the web in making a machine learning model using 1096 historical data on river water levels and weather. In the experiments carried out, we also carried out comparisons using data that the Synthetic Minority Over-Sampling Technique has carried out with Gaussian Noise (SMOGN) process and data that SMOGN has not processed. The results show that using 125 nodes can produce a smaller prediction error value. For experiments using SMOGN, the results of the experiment are still inconsistent, and some observation points produce larger MAE values.