Comparison of Data Driven Models (DDM) for soil moisture retrieval using microwave remote sensing data

This paper aims to explore the use of various DDM methods for soil moisture retrieval, identifying the advantages and disadvantages of each, compare and evaluate the results for further study. The study looks into the advantages and disadvantages of each DDM method, summarizing the Root-Mean-Square-Error (RMSE) to identify soil moisture condition. In this study, Neural Network Model, Fuzzy-Rule Model, Bayesian Model, Multiple Regression Model and Support Vector Machines (SVM) were reviewed. The Neural Network model performed better compared with other models, proven with the lowest number of RMSE. The SVM model also showed high potential, whereas the Bayesian, Multiple Regression and Fuzzy-Rule Based models showed higher RMSE values, which indicate higher difference in accuracy.