BAGofT (version 1.0.0)

testGlmnet: Testing penalized logistic regressions

Description

testGlmnet specifies a penalized logistic regression as the classifier to test. It returns a function that can be taken as the input of ‘testModel’. R package ‘glmnet’ is required.

Usage

testGlmnet(formula, alpha = 1)

Arguments

formula

an object of class "formula" (or one that can be coerced to that class): a symbolic description of the model to test.

alpha

the elasticnet mixing parameter, with \(0\le\alpha\le 1\). The penalty is defined as $$(1-\alpha)/2||\beta||_2^2+\alpha||\beta||_1.$$ alpha=1 is the lasso penalty, and alpha=0 the ridge penalty.

References

Zhang, Ding and Yang (2021) "Is a Classification Procedure Good Enough?-A Goodness-of-Fit Assessment Tool for Classification Learning" arXiv preprint arXiv:1911.03063v2 (2021).

Examples

# NOT RUN {
###################################################
# Generate a sample dataset.
###################################################
# set the random seed
set.seed(20)
# set the number of observations
n <- 200
# set the number of covariates
p <- 20

# generate covariates data
Xdat <- matrix(runif((n*p), -5,5), nrow = n, ncol = p)
colnames(Xdat) <- paste("x", c(1:p), sep = "")

# generate random coefficients
betaVec <- rnorm(6)
# calculate the linear predictor data
lindat <-  3 * (Xdat[,1] < 2 & Xdat[,1] > -2) + -3 * (Xdat[,1] > 2 | Xdat[,1] < -2) +
  0.5 * (Xdat[,2] + Xdat[, 3] + Xdat[,4] + Xdat[, 5])
# calculate the probabilities
pdat <- 1/(1 + exp(-lindat))

# generate the response data
ydat <- sapply(pdat, function(x) rbinom(1, 1, x))

# generate the dataset
dat <- data.frame(y = ydat, Xdat)

###################################################
# Obtain the testing result
###################################################

# 50 percent training set
testRes1 <- BAGofT(testModel = testGlmnet(formula = y~., alpha = 1),
                  data = dat,
                  ne = n*0.5,
                  nsplits = 20,
                  nsim = 40)
# 75 percent training set
testRes2 <- BAGofT(testModel = testGlmnet(formula = y~., alpha = 1),
                   data = dat,
                   ne = n*0.75,
                   nsplits = 20,
                   nsim = 40)
# 90 percent training set
testRes3 <- BAGofT(testModel = testGlmnet(formula = y~., alpha = 1),
                   data = dat,
                   ne = n*0.9,
                   nsplits = 20,
                   nsim = 40)

# print the testing result.
print(c(testRes1$p.value, testRes2$p.value, testRes3$p.value))
# }