tests/testthat/test-mcmc.R

In speedyBBT: Efficient Bayesian Inference for the Bradley--Terry Model

test_that("speedyBBTm produces results within tolerance", {
#Construct covariance matrix
data("forcedMarriage")
expA  <- expm::expm(forcedMarriage$adjacencyMatrix)
sigma <- diag(diag(expA)^-0.5) %*% expA %*% diag(diag(expA)^-0.5)

#Fit model
forcedMarriageModel <- speedyBBTm(outcome = rep(1, length(forcedMarriage$comparisons$win)),
                                  player1 = forcedMarriage$comparisons$win,
                                  player2 = forcedMarriage$comparisons$lost,
                                  player.prior.var = sigma, n.iter = 10000)

forcedMarriageModelMeans <- colMeans(forcedMarriageModel$lambda[-c(1:100), ])

#Read in means
testMeansPath <- test_path("forcedMarriageModelMeans.csv")
testMeans <- read.csv(testMeansPath)

expect_equal(sum(abs(testMeans - forcedMarriageModelMeans))/76, 0, tolerance = 1e-1)

})


test_that("BBTm produces results within tolerance", {
  #Construct covariance matrix
#Fit model
wimbledonModel <- BBTm(outcome = wimbledon$matches$outcome,
                       player1 = wimbledon$matches$winner,
                       player2 = wimbledon$matches$loser,
                       advantage = wimbledon$matches$secondWeek,
                       formula = ~ rank + points, data = wimbledon$players,
                       n.iter = 4000)



wimbledonModelMeans <- colMeans(wimbledonModel$lambda[-c(1:50), ])

#Read in means
testMeansPath <- test_path("wimbledonModelMeans.csv")
testMeans <- read.csv(testMeansPath)

#Compare within tolerance
expect_equal(sum(abs(testMeans - wimbledonModelMeans))/128, 0, tolerance = 1e-1)

})