inst/unitTests/test_SignificanceCalculator.R

In CausalR: Causal network analysis methods

# A test function for functions in Significance Calculator

test_GetNumberOfPositiveAndNegativeEntries <- function() {
    dataList <- matrix(c(1, 2, 3, 4, 5, 6, 7, -1, 1, 1, 0, -1, 1, +1), ncol = 2)
    dataStats <- CausalR:::GetNumberOfPositiveAndNegativeEntries(dataList)
    checkEquals(dataStats[1], 4)
    checkEquals(dataStats[2], 2)
}

test_AnalysePredictionsList <- function() {
    predictionData <- matrix(c(1, 2, 3, 4, 5, 6, 7, -1, 1, 1, 0, -1, -1, +1), ncol = 2)
    predictionDataStats <- CausalR:::AnalysePredictionsList(predictionData, length(predictionData[, 1]))
    checkEquals(predictionDataStats[1], 3)
    checkEquals(predictionDataStats[2], 3)
    checkEquals(predictionDataStats[3], 1)
}

test_GetScoreForNumbersOfCorrectandIncorrectPredictionsProducesCorrectOutput <- function() {
    checkEquals(CausalR:::GetScoreForNumbersOfCorrectandIncorrectPredictions(c(5, 2, 1, 2)), 4)
    checkEquals(CausalR:::GetScoreForNumbersOfCorrectandIncorrectPredictions(c(5, 7, 1, 2)), -1)
}

test_GetWeightForNumbersOfCorrectandIncorrectPredictionsProducesCorrectOutputWhenReturnLogIsFalse <- function() {
    logOfFactorialOfPredictionListStats <- c(log(factorial(3)), log(factorial(2)), log(factorial(0)))
    checkEquals(CausalR:::GetWeightForNumbersOfCorrectandIncorrectPredictions(1, 0, 1, 1, c(3, 2, 0), c(2, 1, 2), logOfFactorialOfPredictionListStats, 
        returnlog = FALSE), 6)
    checkEquals(CausalR:::GetWeightForNumbersOfCorrectandIncorrectPredictions(2, 1, 0, 0, c(3, 2, 0), c(2, 1, 2), logOfFactorialOfPredictionListStats, 
        returnlog = FALSE), 3)
}

test_GetWeightForNumbersOfCorrectandIncorrectPredictionsProducesCorrectOutputWhenReturnLogIsTrue <- function() {
    logOfFactorialOfPredictionListStats <- c(log(factorial(3)), log(factorial(2)), log(factorial(0)))
    checkEquals(CausalR:::GetWeightForNumbersOfCorrectandIncorrectPredictions(1, 0, 1, 1, c(3, 2, 0), c(2, 1, 2), logOfFactorialOfPredictionListStats, 
        returnlog = TRUE), log(6))
    checkEquals(CausalR:::GetWeightForNumbersOfCorrectandIncorrectPredictions(2, 1, 0, 0, c(3, 2, 0), c(2, 1, 2), logOfFactorialOfPredictionListStats, 
        returnlog = TRUE), log(3))
}

test_PopulateTheThreeByThreeContingencyTablePopulatesTableCorrectly <- function() {
    # Test a symmetric case
    predictionDataStats <- c(2, 3, 2)
    experimentalDataStats <- c(2, 3, 2)
    threeByThreeTable <- CausalR:::PopulateTheThreeByThreeContingencyTable(1, 1, 1, 1, predictionDataStats, experimentalDataStats)
    checkEquals(length(threeByThreeTable), 9)
    checkEquals(threeByThreeTable[1], 1)
    checkEquals(threeByThreeTable[2], 1)
    checkEquals(threeByThreeTable[3], 0)
    checkEquals(threeByThreeTable[4], 1)
    checkEquals(threeByThreeTable[5], 1)
    checkEquals(threeByThreeTable[6], 1)
    checkEquals(threeByThreeTable[7], 0)
    checkEquals(threeByThreeTable[8], 1)
    checkEquals(threeByThreeTable[9], 1)
    predictionDataStats2 <- c(4, 3, 2)
    experimentalDataStats2 <- c(2, 3, 4)
    threeByThreeTable2 <- CausalR:::PopulateTheThreeByThreeContingencyTable(1, 1, 1, 1, predictionDataStats2, experimentalDataStats2)
    checkEquals(length(threeByThreeTable), 9)
    checkEquals(threeByThreeTable2[1], 1)
    checkEquals(threeByThreeTable2[2], 1)
    checkEquals(threeByThreeTable2[3], 2)
    checkEquals(threeByThreeTable2[4], 1)
    checkEquals(threeByThreeTable2[5], 1)
    checkEquals(threeByThreeTable2[6], 1)
    checkEquals(threeByThreeTable2[7], 0)
    checkEquals(threeByThreeTable2[8], 1)
    checkEquals(threeByThreeTable2[9], 1)
}

test_ComputePValueFromDistributionTableFromDistributionTableWorksGivenADistributionTable <- function() {
    matrix1 <- matrix(c(1, 2, 3, 0.5, 0.4, 0.1), ncol = 2)
    totalWeights <- sum(matrix1[, 2])
    checkEquals(CausalR:::ComputePValueFromDistributionTable(1, matrix1, totalWeights), 1)
    checkEquals(CausalR:::ComputePValueFromDistributionTable(2, matrix1, totalWeights), 0.5)
    checkEquals(CausalR:::ComputePValueFromDistributionTable(3, matrix1, totalWeights), 0.1)
    checkEquals(CausalR:::ComputePValueFromDistributionTable(4, matrix1, totalWeights), 0)
}

test_ComputePValueFromDistributionTableWorksGivenUnnormalisedDistributionTable <- function() {
    matrix1 <- matrix(c(1, 2, 3, 5, 4, 1), ncol = 2)
    totalWeights <- sum(matrix1[, 2])
    checkEquals(CausalR:::ComputePValueFromDistributionTable(1, matrix1, totalWeights), 1)
    checkEquals(CausalR:::ComputePValueFromDistributionTable(2, matrix1, totalWeights), 0.5)
    checkEquals(CausalR:::ComputePValueFromDistributionTable(3, matrix1, totalWeights), 0.1)
    checkEquals(CausalR:::ComputePValueFromDistributionTable(4, matrix1, totalWeights), 0)
}

test_CalculateSignificanceCallsCalculateSignificanceUsingCubicAlgorithm <- function() {
    # Check that CalculateSignificance calls CalculateSignificanceUsingCubicAlgorithm correctly
    pValue_direct <- CausalR::CalculateSignificanceUsingCubicAlgorithm(5, c(7, 4, 9), c(6, 6, 8), 1e-05)
    pValue_indirect <- CausalR:::CalculateSignificance(5, c(7, 4, 9), c(6, 6, 8), 1e-05, TRUE)
    checkEquals(pValue_direct, pValue_indirect)
}

test_CalculateSignificanceCallsCalculateSignificanceUsingQuarticAlgorithm <- function() {
    # Check that CalculateSignificance calls CalculateSignificanceUsingQuarticAlgorithm correctly
    pValue_direct <- CausalR::CalculateSignificanceUsingQuarticAlgorithm(5, c(7, 4, 9), c(6, 6, 8))
    pValue_indirect <- CausalR:::CalculateSignificance(5, c(7, 4, 9), c(6, 6, 8), useCubicAlgorithm = FALSE)
    checkEquals(pValue_direct, pValue_indirect)
}