v1.0.0-beta.1 release

Version 1.0.0-beta.1 (Pre-release)

Features

• Task-Parameterized Gaussian Mixture Models: Introducing the core functionality of task-parameterized Gaussian mixture models as described in the Calinon paper link.
• Expectation Maximization Algorithm: Implementation of the Expectation Maximization (EM) algorithm with E-Step and M-Step.
• Optimization Criterion: Added the log-likelihood optimization criterion for model evaluation.
• Component Analysis: Ability to fit trajectories and analyze components.

Bug Fixes

• None in this release.

Enhancements

• Improved Initialization: Added options for custom initialization of weights and means.
• Regularization: Incorporated a regularization factor for the empirical covariance matrix.
• Verbose Mode: Added a verbose mode for printing learning statistics.

Documentation

• Updated README.md: Comprehensive documentation with usage examples, explanations of parameters, and detailed class descriptions.
• Docstrings: Improved inline documentation for better code understanding.
• Doxygen: Added Doxygen documentation in latex and html.

Contributors

• Robin Uhrich

This release marks the initial version of the task-parameterized Gaussian mixture models package. We encourage users to explore the features, provide feedback, and contribute to the further development of the project. Refer to the documentation for detailed information on usage and customization.