Prediksi Curah Hujan Menggunakan Jaringan Saraf Tiruan Dengan Metode Backpropagation

Abstract

Rainfall is an essential factor in various sectors such as agriculture, water resource management, and hydrometeorological disaster mitigation. The unpredictable and nonlinear characteristics of rainfall patterns make conventional statistical methods less effective for accurate forecasting. This study aims to develop a rainfall prediction model using an Artificial Neural Network (ANN) with the Backpropagation method implemented in MATLAB, and analyze the model’s accuracy and performance based on historical rainfall data. The constructed network architecture consists of four input neurons, twelve hidden neurons, and one output neuron. The training results show that the monthly rainfall model performs more well. The findings indicate that the Backpropagation-based ANN model effectively predicts monthly rainfall and can be used as a decision-support tool for water resource planning and management in the Makassar–Maros region