rho_mlogit {bcROCsurface}

Fitting disease models via multinomial logistic models

Description

rho_mlogit is used to fit multinomial logistic models to the disease process in the verified subjects.

Usage

rho_mlogit(formula, data, test = FALSE, maxit = 500, trace = FALSE)

Arguments

formula

an object of class "formula": a symbolic description of the model to be fitted.

data

an optional data frame containing the variables in the model.

test

a logical value indicating whether p-values of the regression coefficients should be returned. Default FALSE.

maxit

maximum number of iterations. Default 500.

trace

switch for tracing estimation process. Default FALSE.

Details

In the formula, the response must be a result of pre_data, a factor with three levels, say 1, 2, 3. These levels correspond to three classes of disease status, e.g., non-dieseased, intermediate, diseased. The last class (class 3) is considered as the reference level in multinomal logistic model. In presence of verification bias, the missing (NA) values correspond to non verified subjects.

Value

rho_mlogit returns a list containing the following components:

coeff

a vector of estimated coefficients.

values

fitted values of the model.

Hess

the Hessian of the measure of fit at the estimated coefficients.

D

the disease status vector used.

X

a design model matrix.

formula

the fomular supplied.

References

To Duc, K., Chiogna, M. and Adimari, G. (2016) Bias-corrected methods for estimating the receiver operating characteristic surface of continuous diagnostic tests. Electronic Journal of Statistics, 10, 3063-3113.

See Also

multinom, nnet

Examples

data(EOC)
dise_na <- pre_data(EOC$D, EOC$CA125)
dise_fact_na <- dise_na$dise
out <- rho_mlogit(dise_fact_na ~ CA125 + CA153 + Age, data = EOC,
                  test = TRUE, trace = TRUE)