New Zealand Statistical Association 2024 Conference

Alejandro C. Frery

Victoria University of Wellington

Ordinal Patterns, features and their distribution

This is joint work with Andrea Rey, Juliana Gambini, Magdalena Lucini

In this presentation, we will explore the concept of ordinal patterns and their significance in analysing time series data. Ordinal patterns encode the ranking properties of small vectors, allowing us to transform real-valued time series into sequences of symbols. This transformation simplifies the analysis by focusing on the order of data rather than its actual values. We will discuss the methodology proposed by Bandt and Pompe, which utilizes two key features: permutation entropy and statistical complexity. These features enable us to assess the unpredictability and structural properties of the data, providing valuable insights into the underlying dynamics of complex systems. The application of ordinal patterns spans various fields, including economics, text analysis, neuroscience, and climatology. By examining the marginal and transition properties of these symbols, we can uncover hidden patterns and trends in diverse datasets. Finally, we will highlight recent developments in the statistical properties of ordinal patterns, showcasing their potential for advancing our understanding of complex systems.