Implementation of RMS Prop Algorithm for Flood Prediction Using Rainfall, Air Temperature, Humidity, Wind, and Tides Data R. Hadapiningradja Kusumodestoni, Agus Subhan Akbar, Fandy Indra Pratama, Adi Sucipto, Teguh Tamrin, Gun Sudiryanto
Universitas Islam Nahdlatul Ulama Jepara, Universitas Wahid Hasyim Semarang
Abstract
Climate change will occur in Indonesia in 2024, especially in the Kudus, Demak and Pati areas, there will be major flooding which will cause 20,772 refugees spread across 59 refugee camps, 93,149 people will be affected by the flood, and there will be 4 deaths. This research aims to predict flood-prone areas in the Demak area and its surroundings based on rainfall, air temperature, humidity, wind and tides, using training data in 2023 and test data in 2024 using the RMS Prop algorithm to predict and test accuracy. and determine the RMSE value. This research uses the python-based RMS Prop algorithm method with data used from the meteorology, climatology and ge-ophysics agency for the Class II Maritime Tanjung Emas Semarang meteoro-logical station. The research results obtained in this study were by conducting the experiment 8 times using a number of epochs of 10, 20, 30, 40, 50, 60, 70, 80 which resulted in the best average and lowest RMSE value at a num-ber of epochs of 30 with the research conclusion being accuracy results of 94.4% and an RMSE value of 0.7090, it can be concluded that the RMS Prop algorithm is able to predict floods well based on rainfall, temperature air, humidity, wind, and tides.
