rvl.stat.dist
Class F

java.lang.Object
  |
  +--rvl.stat.dist.F

public class F
extends java.lang.Object

The F distribution (central and noncentral)


Constructor Summary
F()
           
 
Method Summary
static double cdf(double f, double df1, double df2)
           
static double cdf(double f, double df1, double df2, double lambda)
           
static double lambda(double pwr, double df1, double df2, double alpha)
           
static double power(double alpha)
           
static double power(double lambda, double df1, double df2, double alpha)
           
static double power(double effSize, double df1, double df2, double alpha, boolean random)
           
static double powerr(double alpha)
           
static double quantile(double p, double df1, double df2)
           
static double quantile(double p, double df1, double df2, double lambda)
           
static double rocArea(double lambda, double df1, double df2)
           
static double rocArea(double effSize, double df1, double df2, boolean random)
           
static double rocArea(double effSize, double df1, double df2, boolean random, double eps)
           
static double rocArea(double lambda, double df1, double df2, double eps)
           
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

F

public F()
Method Detail

cdf

public static double cdf(double f,
                         double df1,
                         double df2)
Returns:
cdf of the central F distribution at f with d.f. (df1, df2)

cdf

public static double cdf(double f,
                         double df1,
                         double df2,
                         double lambda)
Returns:
cdf of noncentral F distribution at f with d.f. (df1, df2) and noncentrality lambda

quantile

public static double quantile(double p,
                              double df1,
                              double df2,
                              double lambda)
Returns:
pth quantile of the noncentral F distribution with d.f. (df1, df2) and noncentrality lambda

quantile

public static double quantile(double p,
                              double df1,
                              double df2)
Returns:
pth quantile of the central F distribution with d.f. (df1, df2) and noncentrality lambda

power

public static double power(double lambda,
                           double df1,
                           double df2,
                           double alpha)
Returns:
power of fixed-effects F test with size alpha and d.f. (df1, df2) when the effect size (noncentrality) if lambda

power

public static double power(double effSize,
                           double df1,
                           double df2,
                           double alpha,
                           boolean random)
Returns:
power of right-tailed F test with size alpha and d.f. (df1, df2) when the effect size is effSize. random flags whether it is a fixed- or random-effects test. In an ANOVA context, effSize = {EMS(num) - EMS(denom)} / EMS(denom), i.e., EMS(numerator) / EMS(denom) = 1 + effSize.

lambda

public static double lambda(double pwr,
                            double df1,
                            double df2,
                            double alpha)

rocArea

public static double rocArea(double lambda,
                             double df1,
                             double df2,
                             double eps)
Parameters:
eps - Bound on the error of numerical integration
Returns:
area under the ROC curve with ncp lambda and d.f. df1, df2 This is the integral of power(..., alpha) over alpha.

rocArea

public static double rocArea(double lambda,
                             double df1,
                             double df2)
Returns:
rocArea(lambda, df1, df2, 1e-4)

rocArea

public static double rocArea(double effSize,
                             double df1,
                             double df2,
                             boolean random,
                             double eps)
Parameters:
eps - Bound on the error of numerical integration
Returns:
area under the ROC curve with effect size effSize and d.f. df1, df2 This is the integral of power(..., alpha) over alpha.

rocArea

public static double rocArea(double effSize,
                             double df1,
                             double df2,
                             boolean random)
Returns:
rocArea(lambda, df1, df2, 1e-4)

power

public static double power(double alpha)
Returns:
power at alpha for internally stored parameters. This is an auxiliary function used by rocArea, not intended for other uses

powerr

public static double powerr(double alpha)
Returns:
power at alpha for internally stored parameters. This is an auxiliary function used by rocArea, not intended for other uses