NEWS
synergyMixR 1.0.0
Initial Public Release
- Core mixture model implementations: MPCA (Mixture of Probabilistic Component Analyzers) and MFA (Mixture of Factor Analyzers)
- EM algorithm implementations in C++ via Rcpp/RcppArmadillo for performance
- Soft EM with responsibility-weighted updates
- K-means based initialization for cluster assignments
- Comprehensive model selection tools:
- BIC-based grid search over (K, r)
- ICL (Integrated Complete-data Likelihood) with temperature parameter
- Cross-Validation with 1-SE rule
- Bootstrap stability selection for K
- Two-stage selection for stability
- r-selection framework: global, clusterwise, and hybrid modes
- Rotation and flip utilities for synergy alignment
- Complete EMG preprocessing pipeline (DC removal, bandpass, notch, rectification, envelope, normalization)
- Visualization functions for muscle synergy analysis with unified ggplot2 theme
- Simulation utilities for generating mixture data with optional AR(1) temporal noise
- HPC parallel computation support via future/PBS
- Comprehensive testing infrastructure with testthat
- Quickstart vignette and package documentation