DNPROCreg: Direct nonparametric ROC regression modelling
In npROCRegression: Kernel-Based Nonparametric ROC Regression Modelling

DNPROCreg

R Documentation

Direct nonparametric ROC regression modelling

Description

Estimates the covariate-specific ROC curve in the presence of multidimensional covariates by means of the ROC-GAM regression model presented in Rodriguez- Alvarez et al. (2011)

Usage

DNPROCreg(marker, formula.h = ~1, formula.ROC = ~1, group, tag.healthy, data, 
ci.fit = FALSE, test.partial = NULL, newdata = NULL, 
control = controlDNPROCreg(), weights = NULL)

Arguments

`marker`	A character string with the name of the diagnostic test variable.
`formula.h`	Right-hand formula(s) giving the mean and variance model(s) to be fitted in healthy population. Atomic values are also valid, being recycled.
`formula.ROC`	Right-hand formula giving the ROC regression model to be fitted (ROC-GAM model).
`group`	A character string with the name of the variable that distinguishes healthy from diseased individuals.
`tag.healthy`	The value codifying the healthy individuals in the variable `group`.
`data`	Data frame representing the data and containing all needed variables.
`ci.fit`	A logical value. If TRUE, confidence intervals are computed.
`test.partial`	A numeric vector containing the position of the covariate components in the ROC-GAM formula to be tested for a possible effect. If NULL, no test is performed.. If NULL, no test is performed.
`newdata`	A data frame containing the values of the covariate at which predictions are required.
`control`	Output of the `controlDNROCreg()` function.
`weights`	An optional vector of ‘prior weights’ to be used in the fitting process.

Value

As a result, the function DNPROCreg() provides a list with the following components:

`call`	The matched call.
`model`	Data frame containing all variables and observations used in the fitting process.
`fpf`	Set of false positive fractions (FPF) at which the covariate-specific ROC curve has been estimated.
`newdata`	Data frame containing the values of the covariates at which estimates has been obtained.
`pfunctions`	Matrices containing the estimates of each component of the additive predictor of the ROC-GAM. One matrix contains the effects of the covariates, the other the effect of the FPF. Confidence intervals are returned if required).
`coefficients`	Vector of parametric coefficient of the fitted ROC-GAM.
`ROC`	Estimated covariate-specific ROC curve.
`AUC`	Estimated covariate-specific AUC, and corresponding confidence intervals if required.
`pvalue`	If required, p-values are obtained - with two different bootstrap-based tests - for each model component indicated in argument `test.partial` (T2: `L_{2}`-based test; and T1: `L_{1}`-based test). See Rodriguez-Alvarez et al. (2016).

Author(s)

Maria Xose Rodriguez-Alvarez and Javier Roca-Pardinas

References

Rodriguez- Alvarez, M.X., Roca-Pardinas, J. and Cadarso-Suarez, C. (2011). A new flexible direct ROC regression model - Application to the detection of cardiovascular risk factors by anthropometric measures. Computational Statistics and Data Analysis, 55(12), 3257–3270.

Rodriguez- Alvarez, M.X., Roca-Pardinas, J. and Cadarso-Suarez, C. (2016). Bootstrap-based procedures for inference in nonparametric ROC regression analysis. Technical report.

Examples

data(endosim)
# Fit a model including the interaction between age and gender.
m0 <- DNPROCreg(marker = "bmi", formula.h = "~ gender + s(age) + s(age, by = gender)", 
				formula.ROC = "~ gender + s(age) + s(age, by = gender)", 
				group = "idf_status", 
				tag.healthy = 0, 
				data = endosim, 
				control = list(card.P=50, kbin=30, step.p=0.02))
summary(m0)				
plot(m0)

## Not run: 
# For confidence intervals
set.seed(123)
m1 <- DNPROCreg(marker = "bmi", formula.h = "~ gender + s(age) + s(age, by = gender)", 
				formula.ROC = "~ gender + s(age) + s(age, by = gender)", 
				group = "idf_status", 
				tag.healthy = 0, 
				data = endosim, 
				control = list(card.P=50, kbin=30, step.p=0.02),
				ci.fit = TRUE)
summary(m1)
plot(m1)

# For testing the presence of interaccion between age and gender
set.seed(123)
m2 <- DNPROCreg(marker = "bmi", formula.h = "~ gender + s(age) + s(age, by = gender)", 
				formula.ROC = "~ gender + s(age) + s(age, by = gender)", 
				group = "idf_status", 
				tag.healthy = 0, 
				data = endosim, 
				control = list(card.P=50, kbin=30, step.p=0.02),
				test.partial = 3)
summary(m2)
plot(m2)			

## End(Not run)

npROCRegression documentation built on Aug. 31, 2023, 5:15 p.m.