tests/testthat/test-crosspred.R

In ddml: Double/Debiased Machine Learning

sim_dat <- function(nobs) {
  # generate test data
  nobs <- 100
  X <- cbind(1, matrix(rnorm(nobs*39), nobs, 39))
  Z <- matrix(rnorm(nobs*10), nobs, 10) # overidentified
  y <-  X %*% runif(40) + Z %*% c(1, runif(9)) + rnorm(nobs)
  # Organize and return output
  output <- list(D = D, Z = Z, X = X)
  return(output)
}#SIM_DAT

test_that("crosspred computes with a single model", {
  # generate test data
  nobs <- 100
  X <- cbind(1, matrix(rnorm(nobs*39), nobs, 39))
  Z <- matrix(rnorm(nobs*10), nobs, 10) # overidentified
  y <-  X %*% runif(40) + Z %*% c(1, runif(9)) + rnorm(nobs)
  # Define arguments
  learners <- list(what = ols)
  # Compute cross-sample predictions
  crosspred_res <- crosspred(y, X, Z,
                             learners,
                             sample_folds = 3,
                             compute_insample_predictions = T,
                             silent = T)
  # Check output with expectations
  expect_equal(length(crosspred_res$oos_fitted), length(y))
  expect_equal(length(crosspred_res$is_fitted), 3)
})#TEST_THAT

test_that("crosspred computes with ensemble procedures", {
  # generate test data
  nobs <- 100
  X <- cbind(1, matrix(rnorm(nobs*39), nobs, 39))
  Z <- matrix(rnorm(nobs*10), nobs, 10) # overidentified
  y <-  X %*% runif(40) + Z %*% c(1, runif(9)) + rnorm(nobs)
  # Define arguments
  learners <- list(list(fun = ols),
                 list(fun = ols),
                 list(fun = ols))
  # Compute cross-sample predictions
  crosspred_res <- crosspred(y, X, Z,
                             learners,
                             ensemble_type = c("average", "ols",
                                          "nnls1", "nnls",
                                          "singlebest"),
                             cv_folds = 3,
                             sample_folds = 3,
                             compute_insample_predictions = T,
                             silent = T)
  # Check output with expectations
  expect_equal(dim(crosspred_res$oos_fitted), c(length(y), 5))
  expect_equal(length(crosspred_res$is_fitted), 5)
})#TEST_THAT

test_that("crosspred computes with ensemble procedures & custom weights", {
  # generate test data
  nobs <- 100
  X <- cbind(1, matrix(rnorm(nobs*39), nobs, 39))
  Z <- matrix(rnorm(nobs*10), nobs, 10) # overidentified
  y <-  X %*% runif(40) + Z %*% c(1, runif(9)) + rnorm(nobs)
  # Define arguments
  learners <- list(list(fun = ols),
                   list(fun = ols),
                   list(fun = ols))
  # Define custom weights
  custom_ensemble_weights <- diag(1, length(learners))
  colnames(custom_ensemble_weights) <- c("mdl_ols1", "mdl_ols2", "mdl_ols3")
  # Compute cross-sample predictions
  crosspred_res <- crosspred(y, X, Z,
                             learners,
                             ensemble_type = c("average", "ols",
                                               "nnls1", "nnls",
                                               "singlebest"),
                             cv_folds = 3,
                             sample_folds = 3,
                             custom_ensemble_weights = custom_ensemble_weights,
                             compute_insample_predictions = T,
                             silent = T)
  # Check output with expectations
  expect_equal(dim(crosspred_res$oos_fitted), c(length(y), 8))
  expect_equal(length(crosspred_res$is_fitted), 8)
})#TEST_THAT

test_that("crosspred computes with ensemble procedures and sparse matrices", {
  # generate test data
  nobs <- 100
  X <- cbind(1, matrix(rnorm(nobs*39), nobs, 39))
  Z <- matrix(rnorm(nobs*10), nobs, 10) # overidentified
  y <-  X %*% runif(40) + Z %*% c(1, runif(9)) + rnorm(nobs)
  # Define arguments
  learners <- list(list(fun = ols),
                 list(fun = ols))
  # Compute cross-sample predictions
  crosspred_res <- crosspred(y, as(X, "sparseMatrix"),
                             as(Z, "sparseMatrix"),
                             learners,
                             ensemble_type = c("average", "ols",
                                          "nnls1", "nnls",
                                          "singlebest"),
                             cv_folds = 3,
                             sample_folds = 3,
                             compute_insample_predictions = T,
                             silent = T)
  # Check output with expectations
  expect_equal(dim(crosspred_res$oos_fitted), c(length(y), 5))
  expect_equal(length(crosspred_res$is_fitted), 5)
})#TEST_THAT

test_that("crosspred computes auxilliary predictions", {
  # generate test data
  nobs <- 100
  X <- cbind(1, matrix(rnorm(nobs*39), nobs, 39))
  y <-  X %*% runif(40) + rnorm(nobs)
  # Define arguments
  learners <- list(list(fun = ols),
                   list(fun = ols),
                   list(fun = ols))
  # Compute cross-sample and auxilliary predictions
  crosspred_res <- crosspred(y, X,
                             learners = learners,
                             ensemble_type = c("average", "ols",
                                               "nnls1", "nnls",
                                               "singlebest"),
                             cv_folds = 3,
                             sample_folds = 3,
                             silent = T,
                             auxiliary_X = list(X, X, X))
  # Check output with expectations
  expect_equal(dim(crosspred_res$auxiliary_fitted[[1]]), c(length(y), 5))
})#TEST_THAT