A fan-made generator for Dwarnian words, which uses a simple Markov Chain Model based on the words that have appeared in canon.
Dataset used:
Fidelity (Or: How many characters are used to predict the next one?)