influence: Influence Diagnostics for PLreg Objects

In ffqueiroz/PLreg: Power Logit Regression for Modeling Bounded Data

Influence Diagnostics for PLreg Objects

Description

The influence function provides two influence measures and the generalized leverage for power logit regression models.

Usage

influence(model, graph = TRUE, ...)

Arguments

model	fitted model object of class "PLreg".
graph	logical. If graph = TRUE the plots are shown, if graph = FALSE the plots are not shown. Default is graph = TRUE.
...	currently not used.

Value

influence returns a list with three objects:

case.weights	The values of h_{max} eigenvector based on case weights perturbation scheme (see Queiroz and Ferrari (2022)).
totalLI	The total local influence (see Lesaffre and Verbeke (1998))
GL	The diagonal elements of the generalized leverage matrix.

References

Queiroz, F. F. and Ferrari, S. L. P. (2022). Power logit regression for modeling bounded data. arXiv:2202.01697.

See Also

PLreg, residuals.PLreg, envelope, plot.PLreg

Examples

data("Firm")

fitPL <- PLreg(firmcost ~ sizelog + indcost | sizelog + indcost,
              data = Firm, family = "SLASH", zeta = 2.13)

influence_measures = influence(fitPL, graph = FALSE)
plot(influence_measures$case.weights, type = "h", ylim = c(0,1))
plot(influence_measures$totalLI, type = "h", ylim = c(0,6))
plot(Firm$sizelog, influence_measures$GL, pch = "+")

ffqueiroz/PLreg documentation built on March 1, 2023, 2:27 a.m.