How long the singular value decomposed entropy predicts the stock market? — Evidence from the Dow Jones Industrial Average Index

• A new concept of multi-scales singular value decomposition entropy is proposed.
• The predictive power of the entropies to DJIAI is analyzed for different time scales.
• The entropy has short-term (not more than one month) predictive power for DJIAI.

In this paper, a new concept of multi-scales singular value decomposition entropy based on DCCA cross correlation analysis is proposed and its predictive power for the Dow Jones Industrial Average Index is studied. Using Granger causality analysis with different time scales, it is found that, the singular value decomposition entropy has predictive power for the Dow Jones Industrial Average Index for period less than one month, but not for more than one month. This shows how long the singular value decomposition entropy predicts the stock market that extends Caraiani’s result obtained in Caraiani (2014). On the other hand, the result also shows an essential characteristic of stock market as a chaotic dynamic system.