Skip to Main Content
This paper explores the connection between uncertainty and memory effects of time series associated with complex system. Traditionally, information theory based algorithms, such as Shannon entropy and its relatives, are employed as measurements to describe uncertainty quantitatively. This study brings into focus the important role of the long range memory effects on the uncertainty measurements. The method is applicable to arbitrary complex systems. Financial data are investigated as an example. The approach provides important insights into the predictability of a complex system.