- Given the library has been "feature complete" for a year now, let's bump it up to 1.0! :D
- Updated dependencies
- Multiple instances of stop-words are now correctly ignored - thanks @claytongentry!
- The
:stem
option now accepts any stemming function (still defaults tofalse
) - thanks @duijf!
- Implemented
:dets
storage option - Fixed Bernoulli model's float overflow when dealing with large amount of data
- Added
top
option for restricting number of results - Added
encoded_data
option for use with custom storages
- Introduced the storage mechanism (defaults to
:memory
) - Implemented file system storage
- Adjusted the
:smoothing
option
- Implemented Bernoulli model
- Implemented Binarized Multinomial model
- Upgraded the
stemmer
dependency
- Upgraded the
stemmer
dependency
- Integrated
stemmer
, introduced word stemming
- Performance improvements
- Enforce lowercase matching
- Performance improvements
- Documentation fix
- Improved documentation
- Refactored for better readability and tests
- Tweaked configurations
- Tweaked the test suite
- Implemented TF-IDF
- Fixed frequency/weight accumulator
- More precise keyword matching
- Implemented stop words
classify
now returns ordered results- User-configurable options
- Multinomial naive bayes implementation
- Optional keyword weighting