Estimating properties of a homogeneous bounded soil using machine learning models
This work addresses parameter identification in soil science for applications like agriculture or environmental monitoring, but it is incremental as it applies existing ML methods to simulated data.
The study tackled estimating soil properties from water moisture measurements using machine learning models, finding that Support Vector Machines and Neural Networks achieved near-perfect accuracy with minimal errors, though diffusivity predictions were more accurate than hydraulic conductivity ones.
This work focuses on estimating soil properties from water moisture measurements. We consider simulated data generated by solving the initial-boundary value problem governing vertical infiltration in a homogeneous, bounded soil profile, with the usage of the Fokas method. To address the parameter identification problem, which is formulated as a two-output regression task, we explore various machine learning models. The performance of each model is assessed under different data conditions: full, noisy, and limited. Overall, the prediction of diffusivity $D$ tends to be more accurate than that of hydraulic conductivity $K.$ Among the models considered, Support Vector Machines (SVMs) and Neural Networks (NNs) demonstrate the highest robustness, achieving near-perfect accuracy and minimal errors.