A Similarity-based Assisted Grading for Introductory Programming Course

抄録

In an assignment-based introductory programming course, a teacher is required to grade assignments. However, this is generally a boring, labor-intensive task. Assisted grading is an approach to reduce the effort for a manual grading process by automatically classifying student submissions into groups so that a teacher can manually check only representative submissions. In this study, we try an assisted grading using source code similarity. Given a number of manually graded submissions, our method calculates source code similarity between a new submission with them. If there exists a similar submission, we automatically give the same grade as the submission. Otherwise, we ask the teacher to manually grade the submission and use the grade for future submissions. As a preliminary analysis, we have implemented the idea as a simple algorithm and evaluated the reduction in effort using student submissions in a programming course conducted in the department of the authors.

収録刊行物

詳細情報 詳細情報について

問題の指摘

ページトップへ