The motivation behind this project was due too the many hours spent refining features and user stories. I decided if I could predict the variables we were spending so much time refining, then there'd be more time to develop. The idea of this project was to use Machine Learning to predict story points, color, app, and owner of user stories by taking the title and description as an input. The process transforms the documents(title, description) via TF-IDF and uses this to make the appropriate classifications via SVC, and Random Forest classification algorithms.
50% of the project is finished with all 4 of the models yielding a 90% + accuracy. The next part will be generating text from the feature, down to the user story to fully automate the user story process.