WLinfer (version 1.2.0)

Firth_WL: Firth's method for bias correction of estimators

Description

Firth's method is bias correction option included in WL. Firth_WL and Firth_WL_log provides the bias corrected lambda and phi based on Firth's method.

Usage

Firth_WL(y, init)
Firth_WL_log(y, init)

Value

A vector of corrected estimators lambda and phi.

Arguments

y

a numeric vector.

init

a vector of initial values for iterative algorithm designed to solve the modified likelihood equations.

Background

Non-linear equations to be solved are derived in Kim and Jang (2020).

Details

Firth_WL and Firth_WL_log returns a vector of estimates of parameters corrected by Firth's method which uses the modified likelihood equations. In case of weighted Lindley distribution, two non-linear equations should be solved since the solutions are not in closed form. To this end, R package nleqslv is used. To avoid poor local maxima, other estimators like MMEm or MLEc are recommended to be used as initial values.

References

Hyoung-Moon Kim. and Yu-Hyeong Jang. (2020). New Closed-Form Estimators for Weighted Lindley Distribution. , submitted.

Examples

data <- fail_fiber
Firth_WL(data,MMEm_WL(data))
Firth_WL_log(data,MMEm_WL(data))