Title: Accuracy of rainfall prediction using deep learning based on a recurrent neural network with an LSTM layer method

Authors: Rindra Yusianto; Rabei Raad Ali; Pulung Nurtantio Andono; Herwin Suprijono

Addresses: Faculty of Engineering, Department of Industrial Engineering, Universitas Dian Nuswantoro, Semarang, 50131, Central Java, Indonesia ' Department of Computer Engineering Technology, Northern Technical University, Mosul, 41002, Nineveh Governorate, Iraq ' Faculty of Computer Science, Department of Informatics Engineering, Universitas Dian Nuswantoro, Semarang, 50131, Central Java, Indonesia ' Faculty of Engineering, Department of Electrical Engineering, Universitas Dian Nuswantoro, Semarang, 50131, Central Java, Indonesia

Abstract: This study aims to improve the accuracy of rainfall prediction using deep learning on agro-industrial commodity planting land. The variables used are temperature, humidity, and rainfall data as training and test data. The deep learning model used in this study is a recurrent neural network (RNN) with a long-short-term memory (LSTM) layer. The study results show that the RNN model with the LSTM layer in deep learning can predict rainfall based on temperature and humidity data. We found that the accuracy during the training stage using 375 LSTM layers was better with 100 epochs because the result was 91.25% compared to 1000 epochs, only obtaining an accuracy of 89.32%, and 1500 epochs of 87.27%.

Keywords: deep learning; LSTM layer; RNN; recurrent neural network; prediction accuracy; rainfall prediction.

DOI: 10.1504/IJW.2026.153191

International Journal of Water, 2026 Vol.17 No.3, pp.185 - 207

Received: 26 Apr 2025
Accepted: 11 Aug 2025
Published online: 29 Apr 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article