By Thiemen Mussche. The notebook is the main material.
Levenberg-Marquardt is a numerical algorithm specifically made as an improvement to the Gauss-Newton method. It employs Gradient descent to counter some disadvantages of Gauss-Newton. Both algorithms are meant to solve nonlinear least square problems. In the notebook we build the algorithms from the ground up trying to get a more intuitive look at them.
After the theory, the algorithms are implemented and compared on a lowly complex nonlinear function.
