Fitting Nonlinear Equations with the Levenberg–Marquardt Method on Google Earth Engine

Google Earth Engine (GEE) has been widely used to process geospatial data in recent years. Although the current GEE platform includes functions for fitting linear regression models, it does not have function fit nonlinear limiting platform’s capacity and application. To circumvent this limitation, work proposes a general adaptation of Levenberg–Marquardt (LM) method models parallel processing framework its integration into GEE. We compared two commonly methods, LM least square (NLS) methods. found that was superior NLS when we convergence speed, initial value stability, accuracy fitted parameters; therefore, then applied develop platform. further tested by double-logistic equation with global leaf area index (LAI), normalized difference vegetation (NDVI), enhanced (EVI) concluded developed fast, stable, accurate remote sensing data. Given generality algorithm, believe can also be other types equations sorts datasets on