Changes in version 0.1.0                        

Initial Release

Detection Methods

  - Added PELT (Pruned Exact Linear Time) for optimal changepoint
    detection
  - Added CUSUM for cumulative sum based detection
  - Added Binary Segmentation for greedy recursive detection
  - Added Wild Binary Segmentation for robust detection
  - Added BOCPD (Bayesian Online Changepoint Detection)
  - Added Shiryaev-Roberts procedure for sequential detection
  - Added ensemble method combining multiple algorithms

Prior Specifications

  - normal_gamma() for unknown mean and variance
  - normal_known_var() for known variance models
  - normal_wishart() for multivariate data
  - poisson_gamma() for count data
  - geometric_hazard() for constant hazard rate

Features

  - Unified detect_regimes() interface for all methods
  - Online detection with regime_detector() class
  - Uncertainty quantification via bootstrap and Bayesian posteriors
  - Multivariate changepoint detection
  - Comprehensive evaluation metrics
  - ggplot2-based visualization suite
  - Method comparison framework

Documentation

  - Introduction vignette
  - Offline detection vignette
  - Bayesian methods vignette