Sooth
Sooth is a minimal stochastic predictive model. It is used by MegaHAL, a learning chatterbot, and by Typing Simulator, a program that pretends to type text files like a human being.
Getting Started
The basic philisophy is to keep things simple. Sooth can make predictions about events that occur within some context.
A context is just a number; it's up to you to make it significant. It might represent a word, or an array of words, or something else entirely. Sooth doesn't care. Likewise, an event is a number, but doesn't need to represent the same kind of thing as a context. It's perfectly fine for a context to be an array of words and an event to be a single word, which is what MegaHAL does. Or a context could be an array of characters and an event could be a time in milliseconds, which is what TypingSimulator does. Or they could be something else entirely...
Look at the API docs to get up to speed, and at the specs for example usage.
Copyright
Copyright (c) 2018 Jason Hutchens. See UNLICENSE for further details.