WilcoxonRankSumDistribution
javanpst.distributions.tests

Class WilcoxonRankSumDistribution



  • public class WilcoxonRankSumDistributionextends Object
    An implementation of the Wilcoxon Ranks-Sum distribution. Uses a Singleton pattern to ensure only an instance of the class is allowed.
    • Method Detail

      • getInstance

        public static WilcoxonRankSumDistribution getInstance()
        Singleton pattern access to the distribution
        Returns:
        the instance of the distribution
      • getTable

        public static Incomplete3KeyTable getTable()
        Gets the table of the distribution
        Returns:
        Reference to the table
      • computeLeftProbability

        public double computeLeftProbability(int n,                            int m,                            int R)
        Computes left tail probability of the Wilcoxon Ranks-Sum distribution.
        Parameters:
        n - n parameter
        m - m parameter
        R - rank associated
        Returns:
        p-value associated
      • computeRightProbability

        public double computeRightProbability(int n,                             int m,                             int R)
        Computes right tail probability of the Wilcoxon Ranks-Sum distribution.
        Parameters:
        n - n parameter
        m - m parameter
        R - rank associated
        Returns:
        p-value associated
      • findCriticalValue

        public int findCriticalValue(int n,                    int m,                    double alpha)
        Finds the maximum rank whose p-value associated is equal o less than alpha/2 This method is used for computing confidence interval boundaries
        Parameters:
        n - n parameter
        m - m parameter
        alpha - alpha confidence required
        Returns:
        rank found
      • inverseFindCriticalValue

        public double inverseFindCriticalValue(int n,                              int m,                              int u)
        Finds the p-value associated to a rank, provided that it exists This method is used for computing exact confidence for confidence intervals
        Parameters:
        n - n parameter
        m - m parameter
        u - rank selected
        Returns:
        confidence associated

SCaVis 2.0 © jWork.ORG