The distributors of change points in long memory processes
Degree GrantorUniversity of Canterbury
Degree NameBachelor of Science with Honours
In this paper we present the properties of empirical distributions of different statistics (e.g. standard deviation and number of breaks) related to the presence of structural breaks in simulated Fractional Gaussian Noise series with various Hurst parameters. Structural Breaks are detected with Atheoretical Regression Trees, a structural break identification method. The simulation results were applied to four case studies to check whether a Regime Switching or Fractional Gaussian Noise model is more adequate.
- Engineering: Reports