Soil Moisture Prediction of Maize by Combining Support Vector Machine and Chaotic Whale Optimization Algorithm

AbstractSoil moisture (SM) of maize has an extremely important impact on the growth and development of maize. Therefore, predicting the SM of maize will become extremely important, guiding the management of agricultural irrigation and water resources. This paper, therefore, proposes a new hybrid machine learning model (SVM-SWOA), which for the first time incorporates the Whale optimization algorithm (WOA) into a Sinusoidal chaotic map and couples it into a Support Vector Machine (SVM). The WOA will help search for the best parameters to obtain the optimal solution, thus increasing the convergence speed of the algorithm and greatly enhancing the accuracy of the model. This study uses the data from two maize agricultural districts in Iowa, USA, for model creation. Taylor plots and significance tests were used for the data in the study to enable the model to identify the input variables. In order to verify the performance of the model, this paper makes a comprehensive comparison between SVM-SWOA and SVM model. The results show that the coefficient of determination (R2), mean squared error (RMSE) and mean absolute error (MAE) of the new heuristic hybrid model at 60 cm depth were 0.9882, 0.35, and 0.72, respectively, demonstrating the high performance of the SVM-SWOA model in predicting SM. Meta-heuristic algorithms will provide better guidance recommendations for smart agriculture and precision irrigation