Learn R Programming

logistf (version 1.26.0)

flac: FLAC - Firth's logistic regression with added covariate

Description

flac implements Firth's bias-reduced penalized-likelihood logistic regression with added covariate.

Usage

flac(...)

# S3 method for default flac( formula, data, model = TRUE, control, modcontrol, weights, offset, na.action, pl = TRUE, plconf = NULL, ... )

# S3 method for logistf flac(lfobject, data, model = TRUE, ...)

Value

A flac object with components:

coefficients

The coefficients of the parameter in the fitted model.

predict

A vector with the predicted probability of each observation

linear.predictors

A vector with the linear predictor of each observation.

prob

The p-values of the specific parameters

ci.lower

The lower confidence limits of the parameter.

ci.upper

The upper confidence limits of the parameter.

call

The call object.

alpha

The significance level: 0.95

var

The variance-covariance-matrix of the parameters.

loglik

A vector of the (penalized) log-likelihood of the restricted and the full models.

n

The number of observations.

formula

The formula object.

augmented.data

The augmented dataset used

df

The number of degrees of freedom in the model.

method

depending on the fitting method 'Penalized ML' or Standard ML'.} \item{method.ci}{the method in calculating the confidence intervals, i.e. profile likelihood' or `Wald', depending on the argument pl and plconf.

control

a copy of the control parameters.

modcontrol

a copy of the modcontrol parameters.

terms

the model terms (column names of design matrix).

model

if requested (the default), the model frame used.

Arguments

...

Further arguments passed to the method or logistf-call.

formula

A formula object, with the response on the left of the operator, and the model terms on the right. The response must be a vector with 0 and 1 or FALSE and TRUE for the outcome, where the higher value (1 or TRUE) is modeled.

data

A data frame containing the variables in the model.

model

If TRUE the corresponding components of the fit are returned.

control

Controls iteration parameter. Taken from logistf-object when specified. Otherwise default is control= logistf.control().

modcontrol

Controls additional parameter for fitting. Taken from logistf-object when specified. Otherwise default is logistf.mod.control().

weights

specifies case weights. Each line of the input data set is multiplied by the corresponding element of weights

offset

a priori known component to be included in the linear predictor

na.action

a function which indicates what should happen when the data contain NAs

pl

Specifies if confidence intervals and tests should be based on the profile penalized log likelihood (pl=TRUE, the default) or on the Wald method (pl=FALSE).

plconf

specifies the variables (as vector of their indices) for which profile likelihood confidence intervals should be computed. Default is to compute for all variables.

lfobject

A fitted logistf object.

Methods (by class)

  • flac(default): With formula and data

  • flac(logistf): With logistf object

Details

FLAC is a simple modification of Firth's logistic regression which provides average predicted probabilities equal to the observed proportion of events, while preserving the ability to deal with separation. It has been described by Puhr et al (2017).

The modified score equations to estimate coefficients for Firth's logistic regression can be interpreted as score equations for ML estimates for an augmented data set. This data set can be created by complementing each original observation i with two pseudo-observations weighted by \(h_i/2\) with unchanged covariate values and with response values set to \(y=0\) and \(y=1\) respectively. The basic idea of FLAC is to discriminate between original and pseudo-observations in the alternative formulation of Firth's estimation as an iterative data augmentation procedure. The following generic methods are available for ' flac's output object: print, summary, coef, confint, anova, extractAIC, add1, drop1, profile, terms, nobs, predict. Furthermore, forward and backward functions perform convenient variable selection. Note that anova, extractAIC, add1, drop1, forward and backward are based on penalized likelihood ratio tests.

References

Puhr R, Heinze G, Nold M, Lusa L, Geroldinger A (2017). Firth's logistic regression with rare events: accurate effect estimates and predictions? Statistics in Medicine 36: 2302-2317.

See Also

logistf() for Firth's bias-Reduced penalized-likelihood logistic regression.

Examples

Run this code
#With formula and data:
data(sex2)
flac(case ~ age + oc + vic + vicl + vis + dia, sex2)

#With a logistf object:
lf <- logistf(formula = case ~ age + oc + vic + vicl + vis + dia, data = sex2)
flac(lf, data=sex2)

Run the code above in your browser using DataLab