re-write the markov module to be more dynamic, have persistant memory, and a new model for how to build sentences
Markov chain parts of speech, each part of speech having a word list with percentage usages, we lose the context but the parts of speech basically ad-lib us, so it should still be interesting and coherent enough to be fun.
re-write the markov module to be more dynamic, have persistant memory, and a new model for how to build sentences
Markov chain parts of speech, each part of speech having a word list with percentage usages, we lose the context but the parts of speech basically ad-lib us, so it should still be interesting and coherent enough to be fun.
for detecting parts of speech, use a library like https://github.com/dariusk/pos-js