Andrew Pyle | Project | M.S. Information Science | UA Little Rock
This project aims to create a text generator program that will be able to recognizably reproduce the manner of speech of any specific human speaker/writer. It will combine the features of markov chain transition state properties and the prinicples of natural language generation to produce idiosyncratic output based on a training corpus of an individual's writing or speech.
After training on a large corpus, markov chains can be applied to create similar but mostly nonsensical text. This limitation is a major obstacle to the application of markov chains to the field of Natural Language Processing (NLP). This idea overcomes that obstacle.
The project will satisfy the M.S. Project requirement at UA Little Rock. I will be actively working from Fall 2017 through Spring 2018. Check back often for progress!
Progress updated at https://andrew-pyle.github.io/msis-project/