lr_birthdate: Likelihood Ratio for Birth Date

In mispitools: Missing Person Identification Tools

Likelihood Ratio for Birth Date

Description

Computes likelihood ratios (LRs) based on the discrepancy between the actual birth date (ABD) of the missing person and the declared birth date (DBD) of the person of interest. Uses Dirichlet distribution to model category probabilities.

Usage

lr_birthdate(
  ABD = "1976-05-31",
  DBD = "1976-07-15",
  alpha = c(1, 4, 60, 11, 6, 4, 4),
  cuts = c(-120, -30, 30, 120, 240, 360),
  type = 1,
  PrelimData = NULL,
  draw = 500,
  seed = 123
)

Arguments

ABD	Character or Date. Actual birth date of the missing person in "YYYY-MM-DD" format. Default: "1976-05-31".
DBD	Character or Date. Declared birth date of the person of interest in "YYYY-MM-DD" format. Default: "1976-07-15".
alpha	Numeric vector. Alpha parameters for the Dirichlet distribution, typically representing frequencies of solved cases in each discrepancy category. Length should be one more than length of cuts. Default: c(1, 4, 60, 11, 6, 4, 4).
cuts	Numeric vector. Cutoff values (in days) for categorizing the difference between DBD and ABD. Creates length(cuts)+1 categories. Default: c(-120, -30, 30, 120, 240, 360).
type	Integer (1 or 2). Type of search scenario: 1: Open search - MP may not be in database (uses uniform H2) 2: Closed search - MP is in database (uses database frequencies) Default: 1.
PrelimData	Data.frame. Required when type = 2. Contains DBD column for persons of interest in the database. Can be output from sim_poi_prelim.
draw	Integer. Number of Dirichlet samples for probability estimation. Default: 500.
seed	Integer. Random seed for reproducibility. Default: 123.

Details

Categories: The difference between DBD and ABD (in days) is categorized using the cuts vector. Default categories are:

1. < -120 days (DBD more than 4 months before ABD)

2. -120 to -30 days

3. -30 to 30 days (close match)

4. 30 to 120 days

5. 120 to 240 days

6. 240 to 360 days

7. > 360 days (DBD more than 1 year after ABD)

Dirichlet Model: Uses method of moments to estimate category probabilities from Dirichlet samples. The alpha parameter reflects prior knowledge from solved cases about the distribution of birth date discrepancies.

LR Calculation:

• Type 1: LR = P(category | H1) / (1/n_categories)

• Type 2: LR = P(category | H1) / P(category in database)

Value

Numeric. The likelihood ratio for the given birth date discrepancy. Also printed to console.

Deprecation

Soft-deprecated in mispitools 2.0. Generalised by nongenetic_feature (a date feature); the deterministic reference is the Dirichlet mean alpha / sum(alpha) over the discrepancy bins, which the legacy stochastic method-of-moments estimator converges to. The legacy function still works for the 2.0 release-candidate cycle and will be removed afterwards.

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

nongenetic_feature for the unified replacement, sim_lr_prelim for simulating LR distributions, sim_poi_prelim for generating preliminary databases.

Examples

# Type 1: Open search - close match (45 days difference)
lr1 <- lr_birthdate(
  ABD = "1976-05-31",
  DBD = "1976-07-15",
  type = 1,
  seed = 123
)

# Type 1: Open search - larger discrepancy
lr2 <- lr_birthdate(
  ABD = "1976-05-31",
  DBD = "1977-03-15",
  type = 1,
  seed = 123
)

## Not run: 
# Type 2: Closed search with database
# Requires a large database with varied birth dates
db <- sim_poi_prelim(numsims = 1000, seed = 456)
lr3 <- lr_birthdate(
  ABD = "1976-05-31",
  DBD = "1976-07-15",
  type = 2,
  PrelimData = db,
  seed = 123
)

## End(Not run)

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