penalties: likeLTD::penalties

In likeLTD: Tools to Evaluate DNA Profile Evidence

Description

Returns the penalty for the current arguments

Usage

penalties(locusAdjustment, power, dropout, degradation=NULL,
                 rcont=NULL, dropin, locusAdjPenalty=50, dropinPenalty=2,
                 degradationPenalty=50, bemn=-4.35, besd=0.38, ...)

Arguments

locusAdjustment	Locus adjustment for each locus
power	Tvedebrink exponent
dropout	Ignored
degradation	Degradation parameters
rcont	Ignored.
dropin	Dropin rate
locusAdjPenalty	Penalty parameter for the locus adjustments
dropinPenalty	Penalty parameter for the dropin rate
degradationPenalty	Penalty parameter for the degradation parameters
bemn	Mean of the normal distribution used to penalize degradation
besd	Standard deviation of the normal distribution used to penalize degradation
...	Ignored

Details

The penalties are applied if and only if the relevant arguments (locusAdjustment, dropin, degradation, power) are provided. The penalties are as follows:

• dropin:e^{-d*p} where d is the dropin rate and p the associated penalty. The values is normalized to one locus.

• degradation:e^{-p∑ d} where d are the degradation values and p is the associated penalty

• power:dnorm(t, bemn, besd) where t is the Tvedebrink exponent, dnorm is the density of the normal distribution with mean bemn and standard deviation besd

• locusAdjustment:dgamma(a, p, p) where a is the locus adjustment, dgamma is the density of the Gamma distribution with p its shape and rate

Some of these penalties are meant to be applied simultaneously across all loci. Since we want penalties per locus, a normalization p^(1/n) is applied, where p is the penalty and n the number of loci.

Value

An array of penalties per locus

See Also

create.likelihood.vectors, create.likelihood.log, create.likelihood, Objective Functions

likeLTD documentation built on May 1, 2019, 7:58 p.m.