src/test/R/hypergeometricTestCases

# Licensed to the Apache Software Foundation (ASF) under one or more
# contributor license agreements.  See the NOTICE file distributed with
# this work for additional information regarding copyright ownership.
# The ASF licenses this file to You under the Apache License, Version 2.0
# (the "License"); you may not use this file except in compliance with
# the License.  You may obtain a copy of the License at
#
#      http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#
#------------------------------------------------------------------------------
# R source file to validate Hypergeometric distribution tests in
# org.apache.commons.math.distribution.HypergeometricDistributionTest
#
# To run the test, install R, put this file and testFunctions
# into the same directory, launch R from this directory and then enter
# source("<name-of-this-file>")
#
# R functions used
# dhyper(x, m, n, k, log = FALSE) <- density
# phyper(q, m, n, k, lower.tail = TRUE, log.p = FALSE) <- distribution
# qhyper(p, m, n, k, lower.tail = TRUE, log.p = FALSE) <- quantiles
#------------------------------------------------------------------------------
tol <- 1E-6                       # error tolerance for tests
#------------------------------------------------------------------------------
# Function definitions

source("testFunctions")           # utility test functions

# function to verify density computations

verifyDensity <- function(points, expected, good, bad, selected, tol, log = FALSE) {
    rDensityValues <- rep(0, length(points))
    i <- 0
    for (point in points) {
        i <- i + 1
        rDensityValues[i] <- dhyper(point, good, bad, selected, log)
    }
    output <- c("Density test good = ", good, ", bad = ", bad,
                ", selected = ",selected)
    if (assertEquals(expected,rDensityValues,tol,"Density Values")) {
        displayPadded(output, SUCCEEDED, WIDTH)
    } else {
        displayPadded(output, FAILED, WIDTH)
    }
}

# function to verify distribution computations

verifyDistribution <- function(points, expected, good, bad, selected, tol) {
    rDistValues <- rep(0, length(points))
    i <- 0
    for (point in points) {
        i <- i + 1
        rDistValues[i] <- phyper(point, good, bad, selected)
    }
    output <- c("Distribution test good = ", good, ", bad = ",
                 bad, ", selected = ",selected)
    if (assertEquals(expected,rDistValues,tol,"Distribution Values")) {
        displayPadded(output, SUCCEEDED, WIDTH)
    } else {
        displayPadded(output, FAILED, WIDTH)
    }
}

#--------------------------------------------------------------------------
cat("Hypergeometric test cases\n")

good <- 5
bad <- 5
selected <- 5

densityPoints <- c(-1, 0, 1, 2, 3, 4, 5, 10)
densityValues <- c(0, 0.00396825396825397, 0.0992063492063492, 0.396825396825397, 0.396825396825397,
                   0.0992063492063492, 0.00396825396825397, 0)
logDensityValues <- c(-Inf, -5.52942908751142, -2.31055326264322, -0.924258901523332,
                      -0.924258901523332, -2.31055326264322, -5.52942908751142, -Inf)
distributionValues <- c(0, .003968, .103175, .50000, .896825, .996032,
                        1.00000, 1)
#Eliminate p=1 case because it will mess up adjustement below
inverseCumPoints <- c(0, 0.001, 0.010, 0.025, 0.050, 0.100, 0.999,
                      0.990, 0.975, 0.950, 0.900)
inverseCumValues <- c(-1, -1, 0, 0, 0, 0, 4, 3, 3, 3, 3)

verifyDensity(densityPoints, densityValues, good, bad, selected, tol)
verifyDensity(densityPoints, logDensityValues, good, bad, selected, tol, TRUE)
verifyDistribution(densityPoints, distributionValues, good, bad, selected, tol)

i <- 0
rInverseCumValues <- rep(0,length(inverseCumPoints))
for (point in inverseCumPoints) {
  i <- i + 1
  rInverseCumValues[i] <- qhyper(point, good, bad, selected)
}

output <- c("Inverse Distribution test good = ", good, ", bad = ", bad,
            ", selected = ", selected)
# R defines quantiles from the right, need to subtract one
if (assertEquals(inverseCumValues, rInverseCumValues-1, tol,
    "Inverse Dist Values")) {
    displayPadded(output, SUCCEEDED, 80)
} else {
    displayPadded(output, FAILED, 80)
}

# Degenerate cases
good <- 5
bad <- 0
selected <- 3
densityPoints <- c(-1, 0, 1, 3, 10)
densityValues <- c(0, 0, 0, 1, 0)
distributionValues <- c(0, 0, 0, 1, 1)
verifyDensity(densityPoints, densityValues, good, bad, selected, tol)
verifyDistribution(densityPoints, distributionValues, good, bad, selected, tol)

good <- 0
bad <- 5
selected <- 3
densityPoints <- c(-1, 0, 1, 3, 10)
densityValues <- c(0, 1, 0, 0, 0)
distributionValues <- c(0, 1, 1, 1, 1)
verifyDensity(densityPoints, densityValues, good, bad, selected, tol)
verifyDistribution(densityPoints, distributionValues, good, bad, selected, tol)

good <- 3
bad <- 2
selected <- 5
densityPoints <- c(-1, 0, 1, 3, 10)
densityValues <- c(0, 0, 0, 1, 0)
distributionValues <- c(0, 0, 0, 1, 1)
verifyDensity(densityPoints, densityValues, good, bad, selected, tol)
verifyDistribution(densityPoints, distributionValues, good, bad, selected, tol)

displayDashes(WIDTH)