出版社：Canadian Research & Development Center of Sciences and Cultures
摘要：Correlation analysis of financial markets is an important starting point for modern financial theory of financial market risks. Along with the deepening of financial globalization, global financial markets have become more and more interdependent. Correlation analysis of global financial markets has become a hot issue for many scholars. On the basis of an in-depth study of Copula theory, this paper applies the theory to the asymmetric correlation analysis of the global major stock market indexes. First, asymmetric correlations among the selected stock indexes are modeled and detected using the relevant metrics of the Copula function on the logarithmic yield of stock indexes; The detected asymmetric correlations are put together to form a directed acyclic graph. Then, artificial neural networks (ANN) are used as a nonlinear model to predict the nearest future of the target stock index; the prediction accuracy is measured in terms of hit rate and mean square error. Test is done on historical daily data with the results showing that the Copula correlation coefficients are more informative for finding the influential leading markets for the predefined target market better than the traditional linear correlation coefficients. The hit rate of the ANN prediction using the detected leading markets found by Copula correlation coefficients is about 3% to 10% higher than that by the linear correlation coefficients.
其他摘要：Correlation analysis of financial markets is an important starting point for modern financial theory of financial market risks. Along with the deepening of financial globalization, global financial markets have become more and more interdependent. Correlation analysis of global financial markets has become a hot issue for many scholars. On the basis of an in-depth study of Copula theory, this paper applies the theory to the asymmetric correlation analysis of the global major stock market indexes. First, asymmetric correlations among the selected stock indexes are modeled and detected using the relevant metrics of the Copula function on the logarithmic yield of stock indexes; The detected asymmetric correlations are put together to form a directed acyclic graph. Then, artificial neural networks (ANN) are used as a nonlinear model to predict the nearest future of the target stock index; the prediction accuracy is measured in terms of hit rate and mean square error. Test is done on historical daily data with the results showing that the Copula correlation coefficients are more informative for finding the influential leading markets for the predefined target market better than the traditional linear correlation coefficients. The hit rate of the ANN prediction using the detected leading markets found by Copula correlation coefficients is about 3% to 10% higher than that by the linear correlation coefficients.
