A0928
Title: Portfolio optimization through regular vine copula model and computational intelligence method
Authors: Nuttanan Wichitaksorn - Auckland University of Technology (New Zealand) [presenting]
Abstract: A new approach to portfolio investment strategy is introduced by combining the artificial immune system and genetic algorithms in computational intelligence with sentiment analysis. The key component of this strategy is a multivariate regular vine copula-based model with various bivariate copula functions where the marginal model is an intertemporal capital asset pricing with asymmetric exponential generalized autoregressive conditional heteroscedasticity models having a mixture of Gaussian distribution and two generalized Pareto distributions as the innovation. A parallel computing technique is applied to the proposed evolutionary algorithms to accelerate the convergence. In the empirical analysis, two scenarios, with and without the COVID-19 period, are investigated for the dependence structure in the financial markets. In the portfolios, a class of cryptocurrencies is included with the traditional stocks. The proposed model outperforms the benchmark models in terms of risk rewards and diversity indicators.
