## [0.5.0] - 2026-01-24
### Added
- Implemented the Minimal Information Loss algorithm for feature selection and dimension reduction.

## [0.4.0] - 2025-09-13
### Added
- Backend could be selected for BDM. The default is CTM, Lempel–Ziv Complexity (LZC) is an option. Useful for large 2D matrices

## [0.3.1] - 2025-08-29
### Fixed
- Added missing CTM values

## [0.3.0] - 2025-08-29
### Changed
- Replaced the placeholder CTM values with real ones
### Removed
- Removed the incomplete implementation of the `:recursive` boundary condition
- Removed `BDM.Utils.normalize/2`

## [0.2.1] - 2025-08-27
### Added
- Added new simple 3x3 pattern type: `[[1, 0, 0], [0, 0, 0], [0, 0, 0]]` and `[0, 0, 0], [0, 0, 0], [0, 0, 1]]`

## [0.2.0] - 2025-08-26
### Added
- Using the [Nx](https://hexdocs.pm/nx) library to compute BDM for large 2D matrices (images) more efficiently

## [0.1.1] - 2025-08-21
### Fixed
- Fixed README.md 'Usage' section
- Fixed module documentation

## [0.1.0] - 2025-08-12
### Added
- Initial release