Exponential fit code examples using C#
Exponential curves which pass through 0 can be described using two variables. Data can be translated from exponential into linear space, fitted using a linear least squares strategy, then the result translated back into exponential space.
Exponential curves that do not pass through the origin are represented by equations containing 3 unknowns and cannot be fitted using a linear fitter. These advanced equations are best fitted using gradient descent or other methods which seek to minimize the error function. For more, see:
-
SwarmFit - A .NET package for fitting curves to X/Y data using particle swarm optimization
-
Linear Least Squares to Solve Nonlinear Problems - A fantastic YouTube video describing how to use least squares strategy to fit nonlinear functions.
-
C# Helper - A legacy .NET WinForms app which uses gradient descent of the least squares error function to seek the best fit of an exponential curve