IMA: Identification of Multi-author Student Assignment Submissions Using a Data Mining Approach

摘要

In this paper, we describe a novel application of data mining techniques which can be used to identify multi-authorship contained within student submissions. We show that by regarding the pages of the submission as a set of Cascading Style Sheets, CSS type files, which we call author signature styles (ASSs), and accompanying information, it is possible to identify the number of author signature styles contained within the page, or document, irrespective of the number of pages concerned. We also describe how, as a by-product of this work, a set of author signature styles (ASSs) can be created during investigation of each submission and hence be used as a library, containing increasing membership, for comparison with future submissions by the same student. The implications of the use of ASSs for identification of future suspect submissions, and for comparison with future submissions by the same student, are discussed.