PV: Predictive value of the maximum likelihood estimator of the...
In forensim: Statistical tools for the interpretation of forensic DNA mixtures

Description Usage Arguments Value Author(s) References See Also Examples

The PV function implements the predictive value of the maximum likelihood estimator of the number of contributors to a DNA mixture

1	PV(mat, prior)

`mat`	matrix giving the estimates of the conditional probabilities that the maximum likelihood estimator classifies a given stain as a mixture of i contributors given that there are k contributor(s) to the stain. Estimates i must be given in columns for each possible value of the number of contributors given in rows.
`prior`	numeric vector giving the prior probabilities of encountering a mixture of i contributors. `prior` must be of length the number of rows in `mat`.

Vector of the predictive values

Hinda Haned <contact@hindahaned.info>

Haned H., Pene L., Sauvage F., Pontier D., The predictive value of the maximum likelihood estimator of the number of contributors to a DNA mixture, submitted, 2010.

maximum likelihood estimator likestim

# the following examples reproduce some of the calculations appearing
# in the article cited above, for illustrative purpose, the maximum 
#number of contributors is set here to 5 
#matcondi: Table 2 in Haned et al. (2010)
matcondi<-matrix(c(1,rep(0,4),0,0.998,0.005,0,0,0,0.002,0.937,0.067,0,0,0,0.058,
0.805,0.131,rep(0,3),0.127,0.662,rep(0,3),0.001,0.207),ncol=6)
#prior defined by a forensic expert (Table 3 in Haned et al., 2010)
prior1<-c(0.45,0.04,0.30,0.15,0.06)
#uniform prior, for each mixture type, the probability of occurrence is 1/5,
#5 being the threshold for the number of contributors
prior2<-c(rep(1/5,5))
#predictive values for prior1
PV(matcondi,prior1)
#for prior2
PV(matcondi, prior2)