The use case would be to give to digest to an ANTLR variant of, a huge data set of phrases of a language, and hopefully progress to discover a ponderated tree or graph, or some sort of mathematics group , be able to eliminate or proposing for temporary with drawable or some sentences not "near" to the main discovered grammar. The computation would converge when a distance is achieved with out absolute value but Delta around in multiple axes ( verb predicates etc.). The meta grammar would be provided of course to frame the research process.

  • $\begingroup$ The amount of algorithms for grammar induction is huge. Often referenced in the literature are genetic algorithm and neural networks. In most cases a combination of heuristics and data-driven approaches provide the best performance. $\endgroup$ – Manuel Rodriguez Aug 9 '19 at 11:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.