Statistics and Algorithms / Non-Linear Optimization /

Newton-Raphson algorithm

Used by DDV and LDV, Newton–Raphson is a hill-climbing algorithm with takes \(\mathbf{G}=\mathbf{H}^{-1}\), that is, uses the actual inverse Hessian, which isn't easily computable except for certain specific models. It converges very quickly once it is close to the optimum.