tef_penalizedErr: Penalize error for being close to a boundary

In akcochrane/TEfits: Time-evolving model fits

Penalize error for being close to a boundary

Description

TEfit internal. Formally, 'pErr=Err/dbeta(par,dbeta_penalty,dbeta_penalty)' when Err is positive, where par is the (often rate) parameter normalized between 0 [lower bound] and 1 [upper bound]

Usage

tef_penalizedErr(boundedPar, errPar, loBound, upBound, dbeta_penalty = 1.001)

Arguments

boundedPar	Value of the parameter that has bounds.
errPar	Value of the error to be penalized.
loBound	Lower parameter boundary.
upBound	Upper parameter boundary.
dbeta_penalty	The multiplicative error penalty at approximately 5.3 percent away from the boundary. Error increases with increasing proximity to a bound.

akcochrane/TEfits documentation built on June 12, 2025, 11:10 a.m.