Automated extraction of revision events from keystroke data

Abstract Revision plays an important role in writing, and as revisions break down the linearity of writing process, they are crucial describing process dynamics. Keystroke logging analysis have been used to identify made during writing. Previous approaches include manual annotation revisions, building nonlinear S-notations, automated extraction backspace keypresses. However, these time-intensive, vulnerable construct, or restricted. Therefore, this article presents a computational approach automatic full revision events from keystroke logs, including both insertions deletions, well characters typed replace deleted text. Within approach, candidates first automatically extracted, which allows for simplified events. Second, machine learning is detect For this, 7120 were manually annotated dataset keystrokes obtained 65 students conducting task. The results showed that could be predicted with relatively high accuracy. In addition, case study proved easily applied new dataset. To conclude, can beneficial providing insights into