skillScore: Calculate Prediction Skill Scores

In laurieKell/FLCandy: Candidate methods for FLR

Calculate Prediction Skill Scores

Description

Computes threshold-based skill scores including TSS (True Skill Statistic) and AUC (Area Under the Curve) for fishery stock assessment models.

Usage

skillScore(x, y, threshold = NULL, reference = 1)

Arguments

x	Numeric vector of model predictions (e.g., estimated stock biomass)
y	Numeric vector of observed values (e.g., survey biomass index)
threshold	Optional numeric vector of decision thresholds for classification
reference	Reference value for stock status classification (default = 1, typical BMSY threshold)

Value

Data.frame containing optimal threshold and associated metrics:

• ref: Optimal reference threshold

• TP: True positives at threshold

• TN: True negatives at threshold

• FP: False positives at threshold

• FN: False negatives at threshold

• TSS: True Skill Statistic (TPR - FPR)

• TPR: True Positive Rate (sensitivity)

• FPR: False Positive Rate (1 - specificity)

• AUC: Area Under ROC Curve

Examples

# Simulate stock assessment data
obs_status <- rlnorm(100, meanlog = log(1), sdlog = 0.5)
pred_status <- obs_status * exp(rnorm(100, sd = 0.3))
skillScore(pred_status, obs_status > 1, reference = 1)

laurieKell/FLCandy documentation built on April 17, 2025, 5:23 p.m.