Class GARCHModel

  • Direct Known Subclasses:
    GARCH11Model

    public class GARCHModel
    extends Object
    The GARCH(p, q) model takes this form. \[ h_t = \alpha_0 + \sum_{i=1}^{q} (\alpha_i e_{t-i}^2) + \sum_{i=1}^{p} (\beta_i h_{t-i}) \] p is the order of the GARCH terms ht-i; q is the order of the ARCH terms et-i2.
    See Also:
    Wikipedia: GARCH
    • Constructor Summary

      Constructors 
      Constructor Description
      GARCHModel​(double a0, double[] a, double[] b)
      Construct a GARCH model.
      GARCHModel​(GARCHModel that)
      Copy constructor.
    • Constructor Detail

      • GARCHModel

        public GARCHModel​(double a0,
                          double[] a,
                          double[] b)
        Construct a GARCH model.
        Parameters:
        a0 - the constant term
        a - the ARCH coefficients
        b - the GARCH coefficients
      • GARCHModel

        public GARCHModel​(GARCHModel that)
        Copy constructor.
        Parameters:
        that - a GARCH model
    • Method Detail

      • a0

        public double a0()
        Get the constant term.
        Returns:
        the constant term
      • alpha

        public double[] alpha()
        Get the ARCH coefficients.
        Returns:
        the ARCH coefficients; could be null
      • beta

        public double[] beta()
        Get the GARCH coefficients.
        Returns:
        the GARCH coefficients; could be null
      • p

        public int p()
        Get the number of GARCH terms.
        Returns:
        the number of GARCH terms
      • q

        public int q()
        Get the number of ARCH terms.
        Returns:
        the number of ARCH terms
      • maxPQ

        public int maxPQ()
        Get the maximum of the ARCH length or GARCH length.
        Returns:
        max(# ARCH terms, # GARCH terms)
      • var

        public double var()
        Compute the unconditional variance of the GARCH model.
        Returns:
        the unconditional variance
      • sigma2

        public double sigma2​(double[] e2,
                             double[] sigma2_lag)
        Compute the conditional variance based on the past information.
        Parameters:
        e2 - the last q squared observations
        sigma2_lag - the last p conditional variances
        Returns:
        the conditional variance, h(t | Ft)