A module for FCA and Query-Learning
Contains the implementation of algorithms mentioned in
- Queries and Concept Learning (1988): Dana Angluin
- On the Usability of Probably Approximately Correct Implication Bases (2017): Daniel Borchmann, Tom Hanika and Sergei Obiedkov
- Optimizations in computing the Duquenne–Guigues basis of implications (2014): Konstantin Bazhanov, Sergei Obiedkov
ToDo list:
- FCA Part
- PAC Learning
- Implement the dataset sampler
- Complete the approx-equivalent function
- Implement HORN1
- Regex Learning
- Without POS
- Incorporate POS information for potential regex conflicts for different clusters
- Evaluation
- Simple overlapping words evaluation
- Original task metric (would require to implement regex)
- Make automated pipeline for the entire dataset and get cumulative results
- Modulify the code
- Change directory structure
- Add tests
- Add usage documentation