Introduction to GARCH Improved Nelder Mead

GARCH is the short initial for Generalized Autoregressive Conditional Heteroskedasticity and it is the volatility prediction model commonly used in financial industry. GARCH model was first appeared in the work by Danish Economist, Tim Peter Bollerslev in 1986. The 2003 Nobel Prize winner, Robert F Engle also added much contribution for the refinement of GARCH model with Tim’s work. Our GARCH INM predictor took the original method of Nelder Mead for GARCH model building. However, the original Nelder Mead’s method occasionally miss convergence and therefore, we improved the original Nelder Mead method by incorporating intermediate ARMA (Autoregressive Moving Average) step before GARCH step.