leave_one_out: Leave-one-out fragility analysis of evidence contributions

In mispitools: Missing Person Identification Tools

Leave-one-out fragility analysis of evidence contributions

Description

For each step, computes the total weight of evidence with that step removed. Identifies cases where a single step dominates: if fraction[k] is close to 1, the total weight hinges on that single step and the conclusion would be substantially weakened if it were excluded or challenged.

Usage

leave_one_out(per_marker_lrs)

Arguments

per_marker_lrs

Named numeric vector. Per-marker likelihood ratios (positive values).

Value

A data frame with columns

marker

Marker name.

log10_lr

\log_{10} \mathrm{LR}_k, this marker's contribution to the total.

total_without

Total \log_{10} \mathrm{LR} with this marker removed.

fraction

|\log_{10} \mathrm{LR}_k| / \sum_j |\log_{10} \mathrm{LR}_j|, this marker's share of the absolute total.

References

Marsico, F. L. & Egeland, T. (in preparation). Belief dynamics during the investigative process.

See Also

concentration_index, concentration_index_positive.

Examples

lrs <- c(D3S1358 = 5.2, TH01 = 1.8, D21S11 = 120.0, D18S51 = 3.1)
leave_one_out(lrs)

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