An Econometric Approach to Neural Network Model Selection for Financial Time Series Analysis

On Tuesday November 23, Dr. Nikos Thomaidis (University of the Aegean, Chios Island, Greece) presented "An Econometric Approach to Neural Network Model Selection for Financial Time Series Analysis" at the Erasmus University Rotterdam. The purpose of this talk was to present a family of neural network (NN)-GARCH models that jointly parametrise the mean and the variance of the conditional distribution. Dr. Nikos Thomaidis demonstrated how this flexible modelling framework can accommodate most of the stylised facts reported about financial prices or rates of return (nonlinear trends, asymmetric GARCH effects and non-gaussian errors).