Class MarkovCheckNullReference.Result

java.lang.Object
edu.cmu.tetrad.search.MarkovCheckNullReference.Result
Enclosing class:
MarkovCheckNullReference
public static final class MarkovCheckNullReference.Result extends Object
The real-data check statistics and the null distributions, with empirical p-values and component timings.

  • Method Details

    • getDag

      public Graph getDag()
      The DAG actually checked (after any CPDAG-to-DAG conversion).
      Returns:
      The DAG.

    • getSimulator

      public MarkovCheckNullReference.SimulatorType getSimulator()
      The simulator used for the null draws.
      Returns:
      The simulator type.

    • getNumDraws

      public int getNumDraws()
      The number of null draws.
      Returns:
      The number of draws.

    • getRealAdInd

      public double getRealAdInd()
      The real data's Anderson-Darling uniformity p-value over independence facts.
      Returns:
      The value, or NaN if the graph implies no independence facts.

    • getRealFractionDependent

      public double getRealFractionDependent()
      The real data's fraction of independence facts judged dependent.
      Returns:
      The fraction.

    • getNullAdInd

      public double[] getNullAdInd()
      The null draws' Anderson-Darling uniformity p-values.
      Returns:
      A copy of the array, one value per draw.

    • getNullFractionDependent

      public double[] getNullFractionDependent()
      The null draws' fractions of independence facts judged dependent.
      Returns:
      A copy of the array, one value per draw.

    • getEmpiricalP

      public double getEmpiricalP()
      The fraction of null draws with ad_ind at or below the real ad_ind. Small values mean the real data is atypically bad for the null; resolution is 1/numDraws.
      Returns:
      The empirical p, or NaN if the real ad_ind is NaN.

    • getEmpiricalPFraction

      public double getEmpiricalPFraction()
      The fraction of null draws whose rejection fraction is at or above the real one (large fractions are bad, so the tail is on the high side).
      Returns:
      The empirical p, or NaN if the real fraction is NaN.

    • getFitMillis

      public long getFitMillis()
      Time spent fitting the simulator, in milliseconds.
      Returns:
      The time.

    • getSimulateMillis

      public long getSimulateMillis()
      Total time spent simulating the null draws, in milliseconds.
      Returns:
      The time.

    • getCheckMillis

      public long getCheckMillis()
      Total time spent running the Markov checks on the null draws, in milliseconds.
      Returns:
      The time.

    • toString

      public String toString()
      A report in the same format as the py-tetrad prototype.
      Overrides:
      toString in class Object
      Returns:
      The report string.