tfrp

Glass fragments recovered under transfer and persistence

R

tfrp(
  par,
  deffect = TRUE,
  tdist = "nb",
  ymax = 20,
  nsim = 1000,
  seed = NULL,
  hl = 0.02,
  nq = 256
)

The probability of recovering 0, 1, … fragments of glass from a suspect under the transfer, persistence and recovery model of Curran, Triggs, Buckleton, Walsh and Hicks (1998), with uncertain parameters averaged over: the transfer amount and the distance deterministically (by quadrature), the time since the offence exactly, and the loss and recovery rates by nsim Monte Carlo draws (exactly when every rate is fixed).

Parameters

par

The 17 parameters, in order: lam, lsd (transfer amount N(lam, lsd^2), used as its absolute value), d (distance scale), Q (share of high-persistence fragments), l0, u0 (first-hour loss of the others, uniform), ls0, us0 (of the high-persistence ones), lj, uj and lsj, usj (each later hour), lR, uR (share recovered), lt, ut (hours since the offence: fixed when equal) and r (the negative binomial probability of the time).

deffect

TRUE for the distance effect, FALSE without it.

tdist

The law of the time when lt < ut: "nb" (negative binomial with size (lt + ut) / 2 and probability r), "nbtrunc" (that law on floor(lt) to ceiling(ut)) or "unif" (uniform on those hours).

ymax

The largest count reported: a whole number from 0.

nsim

The Monte Carlo draws of the rates: a whole number from 1.

seed

NULL to draw the simulations’ seed from R’s generator (so set.seed() fixes them), or one whole number from 0.

hl, nq

The quadrature’s grid step (1e-4 to 1) and panels (1 to 100000); the defaults resolve the model’s usual settings.

Returns

A list: prob, a data frame of y (0 to ymax), p and se (the Monte Carlo standard error, 0 when exact); cut, the time mass left out and the intensity mass on the grid.

References

Curran, J. M., Triggs, C. M., Buckleton, J. S., Walsh, K. A. J. and Hicks, T. (1998). Assessing transfer probabilities in a Bayesian interpretation of forensic glass evidence. Science and Justice, 38(1), 15-21. doi:10.1016/S1355-0306(98)72068-4

Examples

par <- c(80, 50, 0.5, 0.05, 0.8, 0.9, 0.1, 0.2, 0.4, 0.6, 0.05, 0.1,
         0.5, 0.7, 1, 1, 0.5)
tfrp(par, ymax = 10, nsim = 100, seed = 1)$prob

Generated from the R help page: tfrp in the R reference.