my creation of this predictive text can be run in R studio. This was trained on a corpus of text from political-news sources. The data was used to create n-gram models' very simplified version Katz-Backoff model to generate my first prediction model.
The data corupus cannot be added as it will be around 1.5GB