forensim (version 4.3)

PV: Predictive value of the maximum likelihood estimator of the number of contributors to a DNA mixture

Description

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

Usage

PV(mat, prior)

Arguments

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.

Value

References

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.

See Also

maximum likelihood estimator likestim

Examples

# 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)