Correlations Analysis Between Stock Index Time Serials Based on Reconstructed Phase Space
Gong Weifeng · Computer Technology and Development · 2010
Correlations of stock index is very important for derivative product pricing,risk evaluation,hedge and optimal portfolio choice.Research of correlations of stock index is almost based on linear theory.They think the stock index time serials have a linear characteristics.But in fact,the stock index time serials have a strong non-linear characteristics.Correlations based on linear theory will hava certain limitations.Adopt non-linear chaology to build a multidimensional time serials system by reconstructing gamma space of Shanghai and Shenzhen stock index time serials.Calculate correlations by canonical correlation analysis and achieve correlations between Shanghai stock index and Shenzhen stock index.