autoFlagR 1.0.0
Version 1.0.0 (2025-12-22)
Core Functionality
prep_for_anomaly(): Automated data preprocessing for
anomaly detection
score_anomaly(): Anomaly scoring using Isolation Forest
and Local Outlier Factor
flag_top_anomalies(): Flagging records as anomalous
based on threshold or contamination rate
generate_audit_report(): Generate comprehensive PDF,
HTML, and DOCX audit reports
Benchmarking
extract_benchmark_metrics(): Extract performance
metrics (AUC-ROC, AUC-PR, Top-K Recall)
calculate_benchmark_metrics(): Calculate metrics from
scored data with ground truth
Utilities
get_top_anomalies(): Retrieve top N most anomalous
records
- Feature importance calculation for explaining anomalies
Features
- Support for multiple anomaly detection algorithms (Isolation Forest,
LOF)
- Automated preprocessing (ID handling, scaling, encoding)
- Professional report generation with visualizations
- Benchmarking capabilities for validation studies
- Tidyverse-compatible interface
Documentation
- Comprehensive function documentation
- Getting started vignette
- Healthcare example vignette
- Benchmarking tutorial