transfer: Glass Transfer, Persistence and Recovery Probabilities
In tfer: Forensic Glass Transfer Probabilities

Description Usage Arguments Value Author(s) References Examples

Construct a transfer object to simulate the number of glass fragments recovered given the conditions set by the user.

transfer(
  N = 10000,
  d = 0.5,
  deffect = TRUE,
  lambda = 120,
  Q = 0.05,
  l0 = 0.8,
  u0 = 0.9,
  lstar0 = 0.1,
  ustar0 = 0.15,
  lj = 0.45,
  uj = 0.7,
  lstarj = 0.05,
  ustarj = 0.1,
  lR = 0.5,
  uR = 0.7,
  lt = 1,
  ut = 2,
  r = 0.5,
  timeDist = c("negbin", "cnegbin", "uniform"),
  loop = FALSE
)

`N`	Simulation size
`d`	The breaker's distance from the window
`deffect`	Distance effect. `deffect = TRUE` when distance effect exists. Otherwise `deffect = FALSE`.
`lambda`	The average number of glass fragments transferred to the breaker's clothing.
`Q`	Proportion of high persistence fragments.
`l0`	Lower bound on the percentage of fragments lost in the first hour
`u0`	Upper bound on the percentage of fragments lost in the first hour
`lstar0`	Lower bound on the percentage of high persistence fragments lost in the first hour
`ustar0`	Upper bound on the percentage of high persistence fragments lost in the first hour
`lj`	Lower bound on the percentage of fragments lost in the j'th hour
`uj`	Upper bound on the percentage of fragments lost in the j'th hour
`lstarj`	Lower bound on the percentage of high persistence fragments lost in the j'th hour
`ustarj`	Upper bound on the percentage of high persistence fragments lost in the j'th hour
`lR`	Lower bound on the percentage of fragments expected to be detected in the lab
`uR`	Upper bound on the percentage of fragments expected to be detected in the lab
`lt`	Lower bound on time between commission of crime and apprehension of suspect
`ut`	Upper bound on time between commission of crime and apprehension of suspect
`r`	Probability r in ti ~ NegBinom(t, r)
`timeDist`	the distribution for the random amount of time between the commission of the crime and the apprehension of the suspect. There are three choices `"negbin"`, `"cnegbin"`, and `"uniform"`. Before talking about these it should be noted that if `lt` is equal to `ut` - then there is no randomness in this calculation. If `lt` does not equal `ut`, then the average of these two values is used in the two negative binomial options: `"negbin"` and `"cnegbin"`. The difference betweeen them is that `"cnegbin"` is a constrained negative binomial where the allowable times are constrained to be between `lt` and `ut`. If `"uniform"` is selected, then a uniformly distributed random time between `lt` and `ut` is used in each iteration.
`loop`	if `TRUE` an element by element version of the simulation is used, if `FALSE` then a (mostly) vectorised element version of the simulation is used. The results from the two methods appear to be almost identical - they won't be the same even with the same seed because of the way the random variates are generated. I (James) believe the vectorised version is faster and better. There was also a small mistake which has been corrected in that the initial set of persistent fragments was not being

a list containing:

results: The simulated values of recovered glass fragments
paramList: Input parameters

The returned object has S3 class types tfer and transfer for backwards compatibility

James Curran and TingYu Huang

Curran, J. M., Hicks, T. N. & Buckleton, J. S. (2000). Forensic interpretation of glass evidence. Boca Raton, FL: CRC Press.

Curran, J. M., Triggs, C. M., Buckleton, J. S., Walsh, K. A. J. & Hicks T. N. (January, 1998). Assessing transfer probabilities in a Bayesian interpretation of forensic glass evidence. Science & Justice, 38(1), 15-21.

library(tfer)

## create a transfer object using default arguments
y = transfer()

## probability table
probs = tprob(y)

## extract the probabilities of recovering 8 to 15
## glass fragments given the user-specified arguments
tprob(y, 8:15)

## produce a summary table for a transfer object
summary(y)

## barplot of probabilities (default)
plot(y)
plot(y)

## barplot of transfer frequencies
plot(y, ptype = "f")

## histogram
plot(y, ptype = "h")