tests/testthat/test-regression.R
In NumbersInternational/flipRegression: Estimates standard regression models
context("Regression")
data(bank, package = "flipExampleData")
zformula <- formula("Overall ~ Fees + Interest + Phone + Branch + Online + ATM")
sb <- bank$ID > 100
attr(sb, "label") <- "ID greater than 100"
wgt <- bank$ID
attr(wgt, "label") <- "ID"
bank$dep <- (unclass(bank$Overall) - 1) / 6
missing <- "Exclude cases with missing data"
test_that(missing,
{
z <- as.numeric(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.27732,4))
z <- as.numeric(suppressWarnings(Regression(zformula, data = bank, subset = sb, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.25451,4))
z <- as.numeric(suppressWarnings(Regression(zformula, data = bank, weights = wgt, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.2611546, 4))
z <- as.numeric(suppressWarnings(Regression(zformula, data = bank, weights = wgt, subset = sb, missing = missing))$coef[3])
expect_equal(round(z,4),round(0.2539403,4))
})
missing <- "Dummy variable adjustment"
test_that(missing,
{
z <- as.numeric(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.28768,4))
z <- as.numeric(suppressWarnings(Regression(zformula, data = bank, subset = sb, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.28978,4))
z <- as.numeric(suppressWarnings(Regression(zformula, data = bank, weights = wgt, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.29824, 4))
z <- as.numeric(suppressWarnings(Regression(zformula, data = bank, weights = wgt, subset = sb, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.29795,4))
})
missing <- "Imputation (replace missing values with estimates)"
test_that(missing,
{
if (.Platform$OS.type == "windows")
{
z <- as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, missing = missing)$coef[3])
expect_equal(round(z, 3), 0.316)
z <- as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, subset = sb, missing = missing)$coef[3])
expect_equal(round(z, 3), 0.286)
z <- suppressWarnings(as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, weights = wgt, missing = missing)$coef[3]))
expect_equal(round(z, 3), 0.291)
z <- suppressWarnings(as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, weights = wgt, subset = sb, missing = missing)$coef[3]))
expect_equal(round(z, 3), 0.293)
} else
{
z <- as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, missing = missing)$coef[3])
expect_equal(round(z, 3), 0.305)
z <- as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, subset = sb, missing = missing)$coef[3])
expect_equal(round(z, 3), 0.288)
z <- suppressWarnings(as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, weights = wgt, missing = missing)$coef[3]))
expect_equal(round(z, 3), 0.270)
z <- suppressWarnings(as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, weights = wgt, subset = sb, missing = missing)$coef[3]))
expect_equal(round(z, 3), 0.286)
}
})
missing <- "Multiple imputation"
test_that(missing, {
reg.formula <- Overall ~ Fees + Interest + Phone + Branch + Online + ATM
basic.reg.args <- list(
formula = reg.formula,
data = bank,
missing = missing
)
z <- as.numeric(do.call(Regression, basic.reg.args)$coef[3])
expect_equal(round(z, 3), 0.298)
reg.args <- basic.reg.args
reg.args[["subset"]] <- sb
z <- as.numeric(do.call(Regression, reg.args)$coef[3])
expect_equal(round(z, 3), 0.313)
reg.args <- basic.reg.args
reg.args[["weights"]] <- wgt
z <- as.numeric(do.call(Regression, reg.args)$coef[3])
expect_equal(round(z, 3), 0.312)
reg.args <- basic.reg.args
reg.args[["weights"]] <- wgt
reg.args[["subset"]] <- sb
z <- as.numeric(do.call(Regression, reg.args)$coef[3])
expect_equal(round(z, 3), 0.323)
})
missing <- "Multiple imputation"
test_that("DS-2645 - Entirely Missing predictor", {
bank$Bogus <- rep(NA, nrow(bank))
base.string <- paste0("Overall ~ ", paste0(names(bank)[-which(names(bank) == "Overall")], collapse = " + "))
test.formula <- as.formula(base.string)
expect_warning(Regression(test.formula, data = bank, missing = missing),
"Data has variable(s) that are entirely missing values (all observed values of the variable are missing). These variable(s) have been removed from the analysis: Bogus.", fixed = TRUE)
bank$Nothing <- bank$Bogus
test.formula.with.two.all.missing <- as.formula(paste0(base.string, " + Nothing"))
expect_warning(Regression(test.formula.with.two.all.missing, data = bank, missing = missing),
"Data has variable(s) that are entirely missing values (all observed values of the variable are missing). These variable(s) have been removed from the analysis: Bogus, Nothing.", fixed = TRUE)
bank$Bogus <- NULL
missing.response.formula <- as.formula(gsub("Overall", "Nothing", gsub(" + Bogus", "", base.string, fixed = TRUE)))
expect_error(Regression(missing.response.formula, data = bank),
"Response variable is entirely missing (all observed values of the variable are missing).", fixed = TRUE)
bank$Nothing <- NULL
})
test_that("DS-2645 - Missing predictors and/or interaction", {
bank$Bogus <- rep(NA, nrow(bank))
expect_warning(Regression(Overall ~ Fees + Interest + ATM + Bogus, interaction = Branch, data = bank, missing = missing),
"^Data has variable\\(s\\) that are entirely missing values \\(all observed values of the variable are missing\\). These variable\\(s\\) have been removed from the analysis: Bogus.", perl = TRUE)
bank$Bogus = NULL
bank$Branch2 = rep(NA, nrow(bank))
expect_error(Regression(Overall ~ Fees + Interest + ATM, interaction = Branch2, data = bank, missing = missing),
"Crosstab interaction variable must contain more than one unique value.")
expect_error(Regression(Overall ~ Fees + Interest + ATM, interaction = Branch, data = bank, missing = missing),
NA)
bank$Branch2= NULL
})
#### REDUCE DATA SIZE FOR TESTS WITHOUT NUMERICAL EQUALITY ###
data(bank, package = "flipExampleData")
bank <- bank[sample(nrow(bank), 200), ] # random sample of 200 rows to improve perfomance
zformula <- formula("Overall ~ Fees + Interest + Phone + Branch + Online + ATM")
sb <- bank$ID > 100
attr(sb, "label") <- "ID greater than 100"
wgt <- bank$ID
attr(wgt, "label") <- "ID"
bank$dep <- (unclass(bank$Overall) - 1) / 6
test_that(paste("Alternative ways of passing data in"),
{
type = "Linear"
# no weight, no filter
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, weights = NULL, type = type))
attach(bank)
z1 = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, weights = NULL, type = type))
expect_true(all.equal(z$coef, z1$coef))
detach(bank)
# filter and weight a part of the data frame.
zbank <- cbind(bank, w = wgt, ff = sb)
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = zbank, subset = ff, weights = w, type = type))
attach(zbank)
z1 = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, weights = w, subset = ff, type = type))
detach(zbank)
expect_true(all.equal(z$coefficients, z1$coefficients))
# filter and weight a part of the data frame and are formulas.
zbank <- cbind(bank, w = wgt, ff = sb)
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = zbank, subset = ff == TRUE, weights = w, type = type))
attach(zbank)
z1 = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, weights = w, subset = ff == TRUE, type = type))
detach(zbank)
expect_true(all.equal(z$coefficients, z1$coefficients))
# filter and weight are not part of the data frame.
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, subset = sb, weights = wgt, type = type))
attach(zbank)
z1 = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, subset = sb, weights = wgt, type = type))
detach(zbank)
expect_true(all.equal(z$coefficients, z1$coefficients))
# data frame referenced in formula.
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM,
data = bank, subset = sb, weights = wgt, type = type))
z1 = suppressWarnings(Regression(bank$Overall ~ bank$Fees + bank$Interest + bank$Phone + bank$Branch + bank$Online + bank$ATM,
subset = sb, weights = wgt, type = type))
expect_true(all.equal(z$coefficients, z1$coefficients))
})
test_that(paste("Robust se does something"),
{
type = "Linear"
# no weight, no filter
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, subset = TRUE, weights = NULL, type = type))
zs = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, robust.se = TRUE, subset = TRUE, weights = NULL, type = type))
expect_false(isTRUE(all.equal(z$summary$coefficients[,2], zs$summary$coefficients[,2])))
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, detail = FALSE, subset = TRUE, weights = NULL, type = type))
zs = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, detail = FALSE, robust.se = TRUE, subset = TRUE, weights = NULL, type = type))
expect_false(isTRUE(all.equal(z$summary$coefficients[,2], zs$summary$coefficients[,2])))
expect_error(((z)), NA)
expect_error(((zs)), NA)
# type = "Poisson"
# # no weight, no filter
# z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, subset = TRUE, weights = NULL, type = type))
# zs = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, robust.se = TRUE, subset = TRUE, weights = NULL, type = type))
# expect_false(isTRUE(all.equal(z$summary$coefficients[,2], zs$summary$coefficients[,2])))
# z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, detail = FALSE, subset = TRUE, weights = NULL, type = type))
# zs = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, detail = FALSE, robust.se = TRUE, subset = TRUE, weights = NULL, type = type))
# expect_false(isTRUE(all.equal(z$summary$coefficients[,2], zs$summary$coefficients[,2])))
# expect_error(((z)), NA)
# expect_error(((zs)), NA)
})
type = "Multinomial Logit"
missing = "Multiple imputation"
for(missing in c("Multiple imputation", "Imputation (replace missing values with estimates)",
"Exclude cases with missing data", "Dummy variable adjustment"))
for (type in c("Multinomial Logit", "Linear","Poisson", "Quasi-Poisson","Binary Logit", "Ordered Logit", "NBD"))
test_that(paste("Type by residual", missing, type),
{
# no weight, no filter
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = TRUE, weights = NULL, type = type))
# weight
expect_error(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = sb, weights = NULL, type = type)), NA)
# weight, filter
expect_error(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = TRUE, weights = wgt, type = type)), NA)
# weight, filter
expect_error(z <- suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = sb, weights = wgt, type = type)), NA)
# Prevent pop-ups
mockery::stub(print.Regression, "print.htmlwidget", NULL)
expect_error(suppressWarnings(print(z)), NA)
})
test_that("allEffects works on Regression object",
{
data(cpus, package = "MASS")
z <- Regression(log10(perf) ~ syct+mmin+mmax+cach+chmin+chmax, data = cpus)
zlm <- lm(log10(perf) ~ syct+mmin+mmax+cach+chmin+chmax, data = cpus)
expect_equal(effects::allEffects(z), effects::allEffects(zlm), check.attributes = FALSE)
})
for(missing in c("Multiple imputation", "Imputation (replace missing values with estimates)",
"Exclude cases with missing data", "Dummy variable adjustment"))
for (type in c("Multinomial Logit","Linear","Poisson", "Quasi-Poisson","Binary Logit", "Ordered Logit", "NBD"))
test_that(paste("Stops gracefully with small sample size", missing, type),
{
expect_that(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = wgt > 30000, type = type)), throws_error())
expect_that(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = wgt > 30000, weights = wgt, type = type)), throws_error())
})
test_that("Error due to missing data",
{
missing = "Error if missing data"
expect_error(Regression(zFormula, data = bank, missing = missing), tolerance = 1.0e-8)
})
for(missing in c("Multiple imputation", "Imputation (replace missing values with estimates)",
"Exclude cases with missing data", "Dummy variable adjustment"))
for (type in c("Multinomial Logit", "Linear","Poisson", "Quasi-Poisson","Binary Logit", "Ordered Logit", "NBD"))
for (detail in c(FALSE, TRUE))
test_that(paste("No error", missing, type, "detail =", detail),
{
if (type!= "Multinomial Logit" | detail)
{
# no weight, no filter
expect_error(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = TRUE, detail = detail, weights = NULL, type = type)), NA)
# weight
expect_error(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = sb, detail = detail, weights = NULL, type = type)), NA)
# weight, filter
expect_error(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = TRUE, detail = detail, weights = wgt, type = type)), NA)
# weight, filter
expect_error(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = sb, weights = wgt, detail = detail, type = type)), NA)
} else
expect_true(TRUE)
})
for(missing in c("Imputation (replace missing values with estimates)", "Multiple imputation",
"Exclude cases with missing data", "Dummy variable adjustment"))
for (type in c("Multinomial Logit", "Linear","Poisson", "Quasi-Poisson","Binary Logit", "Ordered Logit", "NBD"))
test_that(paste(type, " save variables"),{
z <- suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, type = type, missing = missing, weights = wgt / 100, subset = sb))
expect_equal(length(predict(z)), 200)
})
for (type in c("Linear","Poisson", "Quasi-Poisson","Binary Logit", "Ordered Logit", "NBD"))
test_that(paste(type, "does not have an error when producing non-detailed outputs"),{
z <- suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, type = type, missing = missing, weights = wgt / 100, subset = sb, detail = FALSE))
# Prevent pop-ups
mockery::stub(print.Regression, "print.htmlwidget", NULL)
expect_error(suppressWarnings(print(z)), NA)
})
test_that("Error if too many dummy predictors",
{
df <- data.frame(outcome = runif(10), pred1 = as.factor(letters[1:10]), pred2 = as.factor(LETTERS[1:10]))
expect_error(Regression(outcome ~ ., data = df),
"There are fewer observations.")
})
test_that("DS-2702: Outcome labels are preserved", {
bank$Overall <- as.factor(bank$Overall)
#bank$Overall[is.na(bank$Overall)] <- 1 # previously label was lost for variables without missing values
attr(bank$Overall, "label") <- "lbl"
result <- suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM,
data = bank, show.labels = TRUE))
expect_equal(result$outcome.label, "lbl")
})
test_that("Ensure output size does not get too large (DS-2518)", {
bank50 <- do.call("rbind", lapply(1:50, function(x) bank))
result <- suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank50))
expect_true(as.numeric(object.size(result)) < 10000000) # less than 10MB
})
test_that("DS-2978: hccm works with a single predictor", {
x <- 1:10
y <- x + rnorm(10)
expect_error(car::hccm(lm(y ~ x - 1)), NA)
})
test_that("Output contains the right class for extension buttons", {
# NOTE: if any of the tests below fail due to class names changing, ALL
# extension buttons in the wiki that refer to this class name should
# be updated with the new class name.
types <- c("Linear", "Binary Logit", "Ordered Logit", "Multinomial Logit",
"Poisson", "Quasi-Poisson", "NBD")
classes <- c("LinearRegression", "BinaryLogitRegression",
"OrderedLogitRegression", "MultinomialLogitRegression",
"PoissonRegression", "QuasiPoissonRegression",
"NBDRegression")
for (i in seq_along(types))
{
result <- suppressWarnings(Regression(Overall ~ Fees + Interest + Phone +
Branch + Online + ATM,
data = bank, type = types[i]))
expect_true(inherits(result, c("Regression", classes[i])))
}
})
NumbersInternational/flipRegression documentation built on March 2, 2024, 10:42 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
context("Regression")
data(bank, package = "flipExampleData")
zformula <- formula("Overall ~ Fees + Interest + Phone + Branch + Online + ATM")
sb <- bank$ID > 100
attr(sb, "label") <- "ID greater than 100"
wgt <- bank$ID
attr(wgt, "label") <- "ID"
bank$dep <- (unclass(bank$Overall) - 1) / 6
missing <- "Exclude cases with missing data"
test_that(missing,
{
z <- as.numeric(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.27732,4))
z <- as.numeric(suppressWarnings(Regression(zformula, data = bank, subset = sb, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.25451,4))
z <- as.numeric(suppressWarnings(Regression(zformula, data = bank, weights = wgt, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.2611546, 4))
z <- as.numeric(suppressWarnings(Regression(zformula, data = bank, weights = wgt, subset = sb, missing = missing))$coef[3])
expect_equal(round(z,4),round(0.2539403,4))
})
missing <- "Dummy variable adjustment"
test_that(missing,
{
z <- as.numeric(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.28768,4))
z <- as.numeric(suppressWarnings(Regression(zformula, data = bank, subset = sb, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.28978,4))
z <- as.numeric(suppressWarnings(Regression(zformula, data = bank, weights = wgt, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.29824, 4))
z <- as.numeric(suppressWarnings(Regression(zformula, data = bank, weights = wgt, subset = sb, missing = missing))$coef[3])
expect_equal(round(z,4), round(0.29795,4))
})
missing <- "Imputation (replace missing values with estimates)"
test_that(missing,
{
if (.Platform$OS.type == "windows")
{
z <- as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, missing = missing)$coef[3])
expect_equal(round(z, 3), 0.316)
z <- as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, subset = sb, missing = missing)$coef[3])
expect_equal(round(z, 3), 0.286)
z <- suppressWarnings(as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, weights = wgt, missing = missing)$coef[3]))
expect_equal(round(z, 3), 0.291)
z <- suppressWarnings(as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, weights = wgt, subset = sb, missing = missing)$coef[3]))
expect_equal(round(z, 3), 0.293)
} else
{
z <- as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, missing = missing)$coef[3])
expect_equal(round(z, 3), 0.305)
z <- as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, subset = sb, missing = missing)$coef[3])
expect_equal(round(z, 3), 0.288)
z <- suppressWarnings(as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, weights = wgt, missing = missing)$coef[3]))
expect_equal(round(z, 3), 0.270)
z <- suppressWarnings(as.numeric(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, weights = wgt, subset = sb, missing = missing)$coef[3]))
expect_equal(round(z, 3), 0.286)
}
})
missing <- "Multiple imputation"
test_that(missing, {
reg.formula <- Overall ~ Fees + Interest + Phone + Branch + Online + ATM
basic.reg.args <- list(
formula = reg.formula,
data = bank,
missing = missing
)
z <- as.numeric(do.call(Regression, basic.reg.args)$coef[3])
expect_equal(round(z, 3), 0.298)
reg.args <- basic.reg.args
reg.args[["subset"]] <- sb
z <- as.numeric(do.call(Regression, reg.args)$coef[3])
expect_equal(round(z, 3), 0.313)
reg.args <- basic.reg.args
reg.args[["weights"]] <- wgt
z <- as.numeric(do.call(Regression, reg.args)$coef[3])
expect_equal(round(z, 3), 0.312)
reg.args <- basic.reg.args
reg.args[["weights"]] <- wgt
reg.args[["subset"]] <- sb
z <- as.numeric(do.call(Regression, reg.args)$coef[3])
expect_equal(round(z, 3), 0.323)
})
missing <- "Multiple imputation"
test_that("DS-2645 - Entirely Missing predictor", {
bank$Bogus <- rep(NA, nrow(bank))
base.string <- paste0("Overall ~ ", paste0(names(bank)[-which(names(bank) == "Overall")], collapse = " + "))
test.formula <- as.formula(base.string)
expect_warning(Regression(test.formula, data = bank, missing = missing),
"Data has variable(s) that are entirely missing values (all observed values of the variable are missing). These variable(s) have been removed from the analysis: Bogus.", fixed = TRUE)
bank$Nothing <- bank$Bogus
test.formula.with.two.all.missing <- as.formula(paste0(base.string, " + Nothing"))
expect_warning(Regression(test.formula.with.two.all.missing, data = bank, missing = missing),
"Data has variable(s) that are entirely missing values (all observed values of the variable are missing). These variable(s) have been removed from the analysis: Bogus, Nothing.", fixed = TRUE)
bank$Bogus <- NULL
missing.response.formula <- as.formula(gsub("Overall", "Nothing", gsub(" + Bogus", "", base.string, fixed = TRUE)))
expect_error(Regression(missing.response.formula, data = bank),
"Response variable is entirely missing (all observed values of the variable are missing).", fixed = TRUE)
bank$Nothing <- NULL
})
test_that("DS-2645 - Missing predictors and/or interaction", {
bank$Bogus <- rep(NA, nrow(bank))
expect_warning(Regression(Overall ~ Fees + Interest + ATM + Bogus, interaction = Branch, data = bank, missing = missing),
"^Data has variable\\(s\\) that are entirely missing values \\(all observed values of the variable are missing\\). These variable\\(s\\) have been removed from the analysis: Bogus.", perl = TRUE)
bank$Bogus = NULL
bank$Branch2 = rep(NA, nrow(bank))
expect_error(Regression(Overall ~ Fees + Interest + ATM, interaction = Branch2, data = bank, missing = missing),
"Crosstab interaction variable must contain more than one unique value.")
expect_error(Regression(Overall ~ Fees + Interest + ATM, interaction = Branch, data = bank, missing = missing),
NA)
bank$Branch2= NULL
})
#### REDUCE DATA SIZE FOR TESTS WITHOUT NUMERICAL EQUALITY ###
data(bank, package = "flipExampleData")
bank <- bank[sample(nrow(bank), 200), ] # random sample of 200 rows to improve perfomance
zformula <- formula("Overall ~ Fees + Interest + Phone + Branch + Online + ATM")
sb <- bank$ID > 100
attr(sb, "label") <- "ID greater than 100"
wgt <- bank$ID
attr(wgt, "label") <- "ID"
bank$dep <- (unclass(bank$Overall) - 1) / 6
test_that(paste("Alternative ways of passing data in"),
{
type = "Linear"
# no weight, no filter
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, weights = NULL, type = type))
attach(bank)
z1 = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, weights = NULL, type = type))
expect_true(all.equal(z$coef, z1$coef))
detach(bank)
# filter and weight a part of the data frame.
zbank <- cbind(bank, w = wgt, ff = sb)
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = zbank, subset = ff, weights = w, type = type))
attach(zbank)
z1 = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, weights = w, subset = ff, type = type))
detach(zbank)
expect_true(all.equal(z$coefficients, z1$coefficients))
# filter and weight a part of the data frame and are formulas.
zbank <- cbind(bank, w = wgt, ff = sb)
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = zbank, subset = ff == TRUE, weights = w, type = type))
attach(zbank)
z1 = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, weights = w, subset = ff == TRUE, type = type))
detach(zbank)
expect_true(all.equal(z$coefficients, z1$coefficients))
# filter and weight are not part of the data frame.
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, subset = sb, weights = wgt, type = type))
attach(zbank)
z1 = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, subset = sb, weights = wgt, type = type))
detach(zbank)
expect_true(all.equal(z$coefficients, z1$coefficients))
# data frame referenced in formula.
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM,
data = bank, subset = sb, weights = wgt, type = type))
z1 = suppressWarnings(Regression(bank$Overall ~ bank$Fees + bank$Interest + bank$Phone + bank$Branch + bank$Online + bank$ATM,
subset = sb, weights = wgt, type = type))
expect_true(all.equal(z$coefficients, z1$coefficients))
})
test_that(paste("Robust se does something"),
{
type = "Linear"
# no weight, no filter
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, subset = TRUE, weights = NULL, type = type))
zs = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, robust.se = TRUE, subset = TRUE, weights = NULL, type = type))
expect_false(isTRUE(all.equal(z$summary$coefficients[,2], zs$summary$coefficients[,2])))
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, detail = FALSE, subset = TRUE, weights = NULL, type = type))
zs = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, detail = FALSE, robust.se = TRUE, subset = TRUE, weights = NULL, type = type))
expect_false(isTRUE(all.equal(z$summary$coefficients[,2], zs$summary$coefficients[,2])))
expect_error(((z)), NA)
expect_error(((zs)), NA)
# type = "Poisson"
# # no weight, no filter
# z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, subset = TRUE, weights = NULL, type = type))
# zs = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, robust.se = TRUE, subset = TRUE, weights = NULL, type = type))
# expect_false(isTRUE(all.equal(z$summary$coefficients[,2], zs$summary$coefficients[,2])))
# z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, detail = FALSE, subset = TRUE, weights = NULL, type = type))
# zs = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, detail = FALSE, robust.se = TRUE, subset = TRUE, weights = NULL, type = type))
# expect_false(isTRUE(all.equal(z$summary$coefficients[,2], zs$summary$coefficients[,2])))
# expect_error(((z)), NA)
# expect_error(((zs)), NA)
})
type = "Multinomial Logit"
missing = "Multiple imputation"
for(missing in c("Multiple imputation", "Imputation (replace missing values with estimates)",
"Exclude cases with missing data", "Dummy variable adjustment"))
for (type in c("Multinomial Logit", "Linear","Poisson", "Quasi-Poisson","Binary Logit", "Ordered Logit", "NBD"))
test_that(paste("Type by residual", missing, type),
{
# no weight, no filter
z = suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = TRUE, weights = NULL, type = type))
# weight
expect_error(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = sb, weights = NULL, type = type)), NA)
# weight, filter
expect_error(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = TRUE, weights = wgt, type = type)), NA)
# weight, filter
expect_error(z <- suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = sb, weights = wgt, type = type)), NA)
# Prevent pop-ups
mockery::stub(print.Regression, "print.htmlwidget", NULL)
expect_error(suppressWarnings(print(z)), NA)
})
test_that("allEffects works on Regression object",
{
data(cpus, package = "MASS")
z <- Regression(log10(perf) ~ syct+mmin+mmax+cach+chmin+chmax, data = cpus)
zlm <- lm(log10(perf) ~ syct+mmin+mmax+cach+chmin+chmax, data = cpus)
expect_equal(effects::allEffects(z), effects::allEffects(zlm), check.attributes = FALSE)
})
for(missing in c("Multiple imputation", "Imputation (replace missing values with estimates)",
"Exclude cases with missing data", "Dummy variable adjustment"))
for (type in c("Multinomial Logit","Linear","Poisson", "Quasi-Poisson","Binary Logit", "Ordered Logit", "NBD"))
test_that(paste("Stops gracefully with small sample size", missing, type),
{
expect_that(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = wgt > 30000, type = type)), throws_error())
expect_that(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = wgt > 30000, weights = wgt, type = type)), throws_error())
})
test_that("Error due to missing data",
{
missing = "Error if missing data"
expect_error(Regression(zFormula, data = bank, missing = missing), tolerance = 1.0e-8)
})
for(missing in c("Multiple imputation", "Imputation (replace missing values with estimates)",
"Exclude cases with missing data", "Dummy variable adjustment"))
for (type in c("Multinomial Logit", "Linear","Poisson", "Quasi-Poisson","Binary Logit", "Ordered Logit", "NBD"))
for (detail in c(FALSE, TRUE))
test_that(paste("No error", missing, type, "detail =", detail),
{
if (type!= "Multinomial Logit" | detail)
{
# no weight, no filter
expect_error(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = TRUE, detail = detail, weights = NULL, type = type)), NA)
# weight
expect_error(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = sb, detail = detail, weights = NULL, type = type)), NA)
# weight, filter
expect_error(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = TRUE, detail = detail, weights = wgt, type = type)), NA)
# weight, filter
expect_error(suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, missing = missing, data = bank, subset = sb, weights = wgt, detail = detail, type = type)), NA)
} else
expect_true(TRUE)
})
for(missing in c("Imputation (replace missing values with estimates)", "Multiple imputation",
"Exclude cases with missing data", "Dummy variable adjustment"))
for (type in c("Multinomial Logit", "Linear","Poisson", "Quasi-Poisson","Binary Logit", "Ordered Logit", "NBD"))
test_that(paste(type, " save variables"),{
z <- suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, type = type, missing = missing, weights = wgt / 100, subset = sb))
expect_equal(length(predict(z)), 200)
})
for (type in c("Linear","Poisson", "Quasi-Poisson","Binary Logit", "Ordered Logit", "NBD"))
test_that(paste(type, "does not have an error when producing non-detailed outputs"),{
z <- suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank, type = type, missing = missing, weights = wgt / 100, subset = sb, detail = FALSE))
# Prevent pop-ups
mockery::stub(print.Regression, "print.htmlwidget", NULL)
expect_error(suppressWarnings(print(z)), NA)
})
test_that("Error if too many dummy predictors",
{
df <- data.frame(outcome = runif(10), pred1 = as.factor(letters[1:10]), pred2 = as.factor(LETTERS[1:10]))
expect_error(Regression(outcome ~ ., data = df),
"There are fewer observations.")
})
test_that("DS-2702: Outcome labels are preserved", {
bank$Overall <- as.factor(bank$Overall)
#bank$Overall[is.na(bank$Overall)] <- 1 # previously label was lost for variables without missing values
attr(bank$Overall, "label") <- "lbl"
result <- suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM,
data = bank, show.labels = TRUE))
expect_equal(result$outcome.label, "lbl")
})
test_that("Ensure output size does not get too large (DS-2518)", {
bank50 <- do.call("rbind", lapply(1:50, function(x) bank))
result <- suppressWarnings(Regression(Overall ~ Fees + Interest + Phone + Branch + Online + ATM, data = bank50))
expect_true(as.numeric(object.size(result)) < 10000000) # less than 10MB
})
test_that("DS-2978: hccm works with a single predictor", {
x <- 1:10
y <- x + rnorm(10)
expect_error(car::hccm(lm(y ~ x - 1)), NA)
})
test_that("Output contains the right class for extension buttons", {
# NOTE: if any of the tests below fail due to class names changing, ALL
# extension buttons in the wiki that refer to this class name should
# be updated with the new class name.
types <- c("Linear", "Binary Logit", "Ordered Logit", "Multinomial Logit",
"Poisson", "Quasi-Poisson", "NBD")
classes <- c("LinearRegression", "BinaryLogitRegression",
"OrderedLogitRegression", "MultinomialLogitRegression",
"PoissonRegression", "QuasiPoissonRegression",
"NBDRegression")
for (i in seq_along(types))
{
result <- suppressWarnings(Regression(Overall ~ Fees + Interest + Phone +
Branch + Online + ATM,
data = bank, type = types[i]))
expect_true(inherits(result, c("Regression", classes[i])))
}
})
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.