ABSTRACT
In soil science, the water content in the soil plays a major role in agriculture and soil chemistry. As the formation of water content depends on various factors, such as field air temperature, field soil temperature, and relative humidity (Rh), accurate prediction of water content was hardly possible without the introduction of any advanced technology. With the advancement of machine learning by various researchers, accurate prediction of water content in the soil became possible. The emphasis of this paper is on the prediction of water content in the soil by utilizing differential evolution (DE) to train in artificial neural network (ANN) to make this prediction possible. A comparison of average root mean square error (RMSE) values and test phase predictions of DE-supported ANN method with other models, such as support vector machine (SVM), multi-layered perceptron (MLP) feed-forward network, Cuckoo Search-supported ANN model, and particle swarm optimization (PSO)-supported ANN model has been reported. Results provided by all the machine-learning models used in this study show that the proposed DE-supported ANN method provides a better solution by reflecting lesser RMSE values than other models in predicting soil water content.
