Add sensitivity, specificity, and ROC curve functions (v0.4.0)

New functions:
- cf_sensitivity(), cf_specificity(), cf_fpr(), cf_tpr(), cf_tnr()
- tr_sensitivity(), tr_specificity(), tr_fpr(), tr_tpr(), tr_tnr()
- cf_roc(), tr_roc() with plot() and as.data.frame() methods

Features:
- Vectorized threshold parameter for efficient computation
- CL, IPW, DR estimators for counterfactual functions
- OM, IPW, DR estimators for transportable functions
- ROC AUC via trapezoidal integration
- Option to compare with naive estimates

Also includes updated vignettes with ROC curve examples.

Author: boyercb

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: cfperformance
 Title: Counterfactual Prediction Model Performance Estimation
-Version: 0.3.0
+Version: 0.4.0
 Authors@R: c(
     person("Christopher", "Boyer", 
            email = "cboyer@hsph.harvard.edu", 

NAMESPACE

@@ -1,18 +1,30 @@
 # Generated by roxygen2: do not edit by hand
 
+S3method(as.data.frame,cf_roc)
+S3method(as.data.frame,tr_roc)
 S3method(coef,cf_performance)
 S3method(coef,tr_performance)
 S3method(confint,cf_performance)
 S3method(confint,tr_performance)
 S3method(plot,cf_auc)
 S3method(plot,cf_calibration)
+S3method(plot,cf_roc)
 S3method(plot,tr_calibration)
+S3method(plot,tr_roc)
 S3method(print,cf_compare)
 S3method(print,cf_cv)
+S3method(print,cf_fpr)
 S3method(print,cf_nuisance)
 S3method(print,cf_performance)
+S3method(print,cf_roc)
+S3method(print,cf_sensitivity)
+S3method(print,cf_specificity)
 S3method(print,ml_learner)
+S3method(print,tr_fpr)
 S3method(print,tr_performance)
+S3method(print,tr_roc)
+S3method(print,tr_sensitivity)
+S3method(print,tr_specificity)
 S3method(summary,cf_compare)
 S3method(summary,cf_cv)
 S3method(summary,cf_performance)
@@ -21,16 +33,30 @@ export(cf_auc)
 export(cf_calibration)
 export(cf_compare)
 export(cf_cv)
+export(cf_fpr)
 export(cf_mse)
+export(cf_roc)
+export(cf_sensitivity)
+export(cf_specificity)
+export(cf_tnr)
+export(cf_tpr)
 export(fit_nuisance)
 export(ml_learner)
 export(tr_auc)
 export(tr_calibration)
+export(tr_fpr)
 export(tr_mse)
+export(tr_roc)
+export(tr_sensitivity)
+export(tr_specificity)
+export(tr_tnr)
+export(tr_tpr)
 importFrom(grDevices,rgb)
 importFrom(graphics,abline)
 importFrom(graphics,hist)
 importFrom(graphics,layout)
+importFrom(graphics,legend)
+importFrom(graphics,lines)
 importFrom(graphics,par)
 importFrom(graphics,polygon)
 importFrom(rlang,.data)

NEWS.md

@@ -1,3 +1,42 @@
+# cfperformance 0.4.0
+
+Adds sensitivity, specificity, and ROC curve functions for both counterfactual
+and transportability settings.
+
+## New Features
+
+### Counterfactual Sensitivity/Specificity
+* `cf_sensitivity()` - Counterfactual sensitivity (true positive rate)
+* `cf_specificity()` - Counterfactual specificity (true negative rate)
+* `cf_fpr()` - Counterfactual false positive rate (1 - specificity)
+* `cf_tpr()` - Alias for `cf_sensitivity()`
+* `cf_tnr()` - Alias for `cf_specificity()`
+* Supports CL, IPW, DR, and naive estimators
+* Vectorized threshold parameter for efficient ROC curve computation
+
+### Transportable Sensitivity/Specificity
+* `tr_sensitivity()` - Transportable sensitivity for target population
+* `tr_specificity()` - Transportable specificity for target population
+* `tr_fpr()` - Transportable false positive rate
+* `tr_tpr()` - Alias for `tr_sensitivity()`
+* `tr_tnr()` - Alias for `tr_specificity()`
+* Supports OM, IPW, DR, and naive estimators
+* Works with both "transport" and "joint" analysis types
+
+### ROC Curves
+* `tr_roc()` - Compute transportable ROC curve in target population
+* `cf_roc()` - Compute counterfactual ROC curve
+* `plot.tr_roc()` / `plot.cf_roc()` - Plot ROC curves with AUC in legend
+* `as.data.frame.tr_roc()` / `as.data.frame.cf_roc()` - Convert to data frame for ggplot2
+* AUC computed via trapezoidal integration
+* Option to include naive ROC curve for comparison
+
+### Documentation
+* Updated introduction vignette with ROC curve examples
+* Updated transportability vignette with ROC curve examples
+
+---
+
 # cfperformance 0.3.0
 
 Adds machine learning integration for flexible nuisance model estimation with