Symbolic Time Series Analysis: A Systematic Review with Entropy-Based Applications in Finance

This paper surveys symbolic encoding procedures that can be successfully utilized in various entropy-based applications. The existing studies indicate several important advantages of the symbolic time series analysis (STSA) in finance and economics, specifically in assessing informational content of financial time series. Data symbolization comprises the conversion of a data series of many different possible values into a symbol series of only a few fixed values. The STSA procedures allow for capturing dynamic time-varying patterns of successive values in financial time series. Discretization techniques can reduce the noise and effectively filter the data. Particularly, they are robust to outliers. Moreover, symbolic encoding of information forms the basis for the Shannon’s mathematical theory of communication and the seminal concept of information entropy.