Version 1.1.0

• Some code optimization, trying to analyze words first with the assumption that no unknown morphemes are involved, if the PoS is already determined
• Adding unseen possibilities to cached observed values for words that are seen less than 10 times in training data
• Slight change in format of training data
• Corrections in German training data.

Version 1.0.0

• Lemmatization is now learned from training data and language independet to some degree
• Training data now have one more column: lemma and stem are both needed
• Scripts for generating training data for English from Brown-Corpus added
• Scripts for generating training data for Dutch from Sonar-Corpus added
• Structure of the project changed with subfolders for creation of training data

Release 0.2.1

Some improvements in creation of training data. Results can be better in some cases but also worse in other cases.

Release 0.2.0

Version from febraury 2021 with many code optimizations that make the analyses significantly faster.

Original Release

Original version of the project. A few bugs solved. Seems to run without problems.