glmfitmiss (version 2.1.0)

logRegMAR: Fitting binary regression with missing categorical covariates using new likelihood based method

Description

This function allows users to fit logistic regression models with incomplete predictors that are categorical. The model is fitted using a new likelihood-based method, which ensures reliable parameter estimation even when dealing with missing data. For more information on the underlying methodology, please refer to Pradhan, Nychka, and Bandyopadhyay (2024).

Usage

logRegMAR(formula, data, conflev = 0.95, correctn = TRUE, verbose = TRUE)

Value

return the logistic regression estimates

Arguments

formula

A formula expression as for regression models, of the form response ~ predictors. The response should be a numeric binary variable with missing values, and predictors can be any variables. A predictor with categorical values with missing can be used in the model. See the documentation of formula for other details.

data

Input data for fitting the model

conflev

Confidence level, the default is 0.95

correctn

a TRUE or FALSE value, by default it is TRUE.

verbose

a TRUE or FALSE value, default is verbose = TRUE

References

Firth, D. (1993). Bias reduction of maximum likelihood estimates, Biometrika, 80, 27-38. doi:10.2307/2336755.

Kosmidis, I., Firth, D. (2021). Jeffreys-prior penalty, finiteness and shrinkage in binomial-response generalized linear models. Biometrika, 108, 71-82. doi:10.1093/biomet/asaa052.

Pradhan, V., Nychka, D. and Bandyopadhyay, S. (2025). Bridging Gaps in Logistic Regression: Tackling Missing Categorical Covariates with a New Likelihood Method (to be submitted).

Pradhan, V., Nychka, D. and Bandyopadhyay, S. (2025). glmFitMiss: Binary Regression with Missing Data in R (to be submitted)

Examples

# \donttest{
# -----------------Example 1: Metastatic Melanoma --------------------------

est1 <- logRegMAR (failcens ~ size+type+nodal+age+sex+trt,
                   data = metastmelanoma, conflev = 0.95, correctn = FALSE)

est1
# -----------------Bias reduced estimates due to Firth (1993) --------------
est2 <- logRegMAR (failcens ~ size+type+nodal+age+sex+trt,
                   data = metastmelanoma, conflev = 0.95, correctn = TRUE)

est2
# -----------------Bias reduced estimates due to Firth (1993) --------------
est2 <- logRegMAR (CaseCntrl ~ Numnill+Numsleep+Smoke+Set+Reftime,
                   data=meningitis, conflev = 0.95, correctn = TRUE)
est2
# }