Parameter free Optimization algorithm for Garch models

12:10 am - 12:30 pm 23 Wednesday

Track

Set Oriented Numerics

In finance, GARCH models are widely used, for example to identify trend, risk and variation in the markets. Nevertless, the choice of the optimal parameters is a hard task. Because of this, in most of cases methods fail and solutions are not good enough.The optimization algorithm presented in this work is a memetic algorithm, which uses two techniques, differential evolution coupled with mathematical programing techniques. The first one is used to find an initial value, and mathematical programing is use to guarantee convergence.We present numerical results on a benchmark related to time series and GARCH models.

Speaker:
Benjamin Perea