To determine the originality of an individual document, a computing system assigns different weights to a plurality of revision events. Each revision event describes a time spent performing a modification to an individual document and a modification type. The weight assigned to each revision event is based on at least the modification type of the revision event. The computing system uses a machine learning model to determine the originality of the individual document based on the plurality of revision events and the different weights assigned to the plurality of revision events. The computing system trains the machine learning model using the plurality of revision events and the determined originality of the individual document.