sim_lr_prelim: Simulate Likelihood Ratios from Preliminary Investigation...

In mispitools: Missing Person Identification Tools

Simulate Likelihood Ratios from Preliminary Investigation Data

Description

Simulates likelihood ratio (LR) distributions based on non-genetic (preliminary investigation) data such as sex, age, region, height, or birth date. This function generates expected LR distributions under both hypotheses:

• H1 (Related): The unidentified person IS the missing person

• H2 (Unrelated): The unidentified person is NOT the missing person

Usage

sim_lr_prelim(
  vartype,
  numsims = 1000,
  seed = 123,
  int = 5,
  ErrorRate = 0.05,
  alphaBdate = c(1, 4, 60, 11, 6, 4, 4),
  numReg = 6,
  MP = NULL,
  database = NULL,
  cuts = c(-120, -30, 30, 120, 240, 360)
)

Arguments

vartype	Character. Type of preliminary investigation variable. Options: "sex", "region", "age", "height", "birthdate".
numsims	Integer. Number of simulations to perform. Default: 1000.
seed	Integer. Random seed for reproducibility. Default: 123.
int	Numeric. Interval parameter for "age" and "height" variables. Defines the estimation range (e.g., if MP age is 55 and int is 10, the range is 45-65). Default: 5.
ErrorRate	Numeric (0-1). Error rate for observations. Default: 0.05.
alphaBdate	Numeric vector. Alpha parameters for Dirichlet distribution used in birthdate LR calculations. Usually frequencies of solved cases in each category. Default: c(1, 4, 60, 11, 6, 4, 4).
numReg	Integer. Number of regions in the case (for "region" variable). Default: 6.
MP	Value of the MP's characteristic for closed search. If NULL, open search is performed. For "sex": "F" or "M"; for "age"/"height": numeric; for "birthdate": date string; for "region": region ID. Default: NULL.
database	Data frame. Database of POIs for closed search (when MP is not NULL). Should have columns matching the variable type (e.g., "Sex", "Age", "Height", "Region", "DBD"). Can be output from sim_poi_prelim.
cuts	Numeric vector. Cutoff values for birthdate categories. Days difference between declared and actual birth dates. Default: c(-120, -30, 30, 120, 240, 360).

Details

Open Search (MP = NULL): Used when it's unknown whether the MP is in the database. LR is computed using general population frequencies as the denominator.

Closed Search (MP specified): Used when comparing a specific MP against a database. LR denominator uses frequencies from the actual database.

Variable-specific calculations:

• sex: Binary match/mismatch with error rate

• region: Match against numReg possible regions

• age/height: Match if within +/- int of MP value

• birthdate: Dirichlet-based probability for date discrepancy

Value

A data.frame with two columns:

• Unrelated: LR values simulated under H2 (POI is not MP)

• Related: LR values simulated under H1 (POI is MP)

Each column contains numsims values.

References

Marsico FL, et al. (2023). "Likelihood ratios for non-genetic evidence in missing person cases." Forensic Science International: Genetics, 66, 102891. \Sexpr[results=rd]{tools:::Rd_expr_doi("10.1016/j.fsigen.2023.102891")}

See Also

sim_lr_genetic for genetic LR simulations, lr_combine for combining LRs from different sources, sim_poi_prelim for creating preliminary databases.

Examples

# Open search for sex variable
lr_sex <- sim_lr_prelim("sex", numsims = 500, seed = 123)
head(lr_sex)

# Check distribution
summary(log10(lr_sex$Related))
summary(log10(lr_sex$Unrelated))

# Visualize
plot_lr_distribution(lr_sex)

# Closed search with database
db <- sim_poi_prelim(numsims = 100, seed = 456)
lr_sex_closed <- sim_lr_prelim(
  "sex",
  numsims = 500,
  MP = "F",
  database = db
)

# Age variable
lr_age <- sim_lr_prelim("age", numsims = 500, int = 10)

mispitools documentation built on Aug. 26, 2026, 1:08 a.m.