statLM 0.0.2

Language models to predict words

statLM (Statistical Language Models) is a library for classical as well as modern language models.

Example Usage

Train a language model and make predictions based on queries i.e. test data.

corpus = ["let us see were this project leads us",
            "we are having great fun so far",
            "we are actively developing",
            "it is getting tougher but it is still fun",
            "this project teaches us how to construct test cases"]

sb = StupidBackoff(n_max=3, alpha=0.4)
# fit model on corpus
sb.fit( corpus )
# make predictions
queries = ["let us see were that project", "how many options"]
sb.predict(queries)

In Progress

• more language models

• improve efficiency of ngram comparisons

• construct CD/CI tests via github action

• add type checking

Copyright

Copyright (C) 2020 statLM Raphael Redmer

For license information, see LICENSE.txt.