Welcome to Biston.
Biston is the third iteration in the Calliope Sentiment Parser Series (CSPS). It is open-source and freely available for use and contributions. Development was put on hold in January 2016 until further interest develops.
Publication
Feel free to download our report here: Biston report
What does Biston do?
Biston uses Bayesian methods to model the polarity of Yelp reviews and estimate word sentiment. It is coupled with a machine-learning model that predicts usefullness of the text. Combining these two metrics together allows companies to quickly see points that affect customer satisfaction, plus it filters out reviews that contain little helpful information.
Contributors
This project is the work of Brigham Young University, done as coursework for the Natural Language Processing course (LING581). You can contact us at thyer@byu.edu or through GitHub.