Comparing ordinal pattern-based short-term predictions to alternative forecasting methods

In this work, the feasibility of using ordinal patterns for short-term prediction is analysed. The accuracy of ordinal pattern-based predictions is compared to the accuracy of the linear predictor method. These comparisons are then used to conclude under which assumptions and conditions the predictions made by the two methods are similar. The considered stochastic processes are: moving averages, autoregressive, and fractional Brownian motion. Additionally, real-world data sets are used to analyze how the prediction methods perform. Keywords: Ordinal patterns, Predictions, Linear predictors, Fractional Brownian motion, Fractional Gaussian noise, Hurst parameter, Moving average processes, Autoregressive processes, Real-world data