Predicting Soil Moisture Levels with Long Short-Term Memory: A Deep Learning Approach Integrated with Internet of Things Data

Water is essential for plant growth, so it is crucial to adjust watering practices to the specific needs of the plant. Factors that affect the watering process include air temperature and soil moisture. This research aims to predict soil moisture levels. Technologies are needed to support this research in obtaining accurate data on soil moisture levels. This research uses Long Short-Term Memory (LSTM) integrated with Internet of Things (IoT) technology. The data will then be processed and analyzed using a quantitative methodology, focusing on the development and application of the Long Short-Term Memory (LSTM) model. Several prediction steps (n) will be analyzed, specifically at steps 6, 12, 18, and 24. We present model performance metrics, including Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination (R2), in both graphical and tabular formats. The results illustrate that the LSTM model achieved the best performance at n = 12, with RMSE = 0.4683 and R2 = 0.8970, indicating a high degree of correlation between predicted and actual values and a low prediction error. This method is expected to improve accuracy and efficiency in predicting soil moisture levels, support decision-making for irrigation strategies, and raise awareness about sustainable water use in agriculture.