Document Difficulty Prediction (2016)

Undergraduate: Duri Long

Faculty Advisor: Prasun Dewan
Department: Computer Science

---

Many students face difficulty when writing documents due to various reasons such as language barriers, content misunderstanding, or lack of formal writing education. While some resources do exist for students who need help with their writing, many are often too shy or too busy to visit a writing center or speak with a professor during office hours. By creating a tool that predicts when students composing documents are facing difficulty and then connects them with help, this problem can be remedied. This thesis focuses on discussing the creation of such a tool and the results obtained from it in a user study. The difficulty prediction tool developed for this thesis project builds on previous research that makes similar predictions for Computer Science students who are writing programs. In the document version of the tool, which consists of a Google Docs plugin, user commands (such as insert, delete, scroll, or highlight) are mapped to command types and then aggregated. The percentage of aggregate commands of a certain type is used to make predictions, which the student can either confirm or correct. The user study, conducted on college students from a variety of different majors, collected data about where the algorithm succeeded and failed in making predictions. Using a graphical analyzer tool developed by the author, the data was evaluated and conclusions were drawn as to how the algorithm could be improved.