Work in progress - A novel methodology to reduce instructors' and students' psychological burdens in source code plagiarism detection

Detecting plagiarism in programming classes is a laborious task for instructors. Many methods have been proposed to reduce the physical strain on them, most of which measure algorithmic or structural similarity between a pair of source codes and are accurate in finding similar source codes. However,...

Full description

Saved in:
Bibliographic Details
Main Authors: Ohno, A, Murao, H
Format: Conference Proceeding
Language:English
Subjects:
Coding style
Encoding
Feature extraction
Hidden Markov models
In-class source codes
Plagiarism
Probability
Programming
Psychological burden
Psychology
Source code plagiarism
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Detecting plagiarism in programming classes is a laborious task for instructors. Many methods have been proposed to reduce the physical strain on them, most of which measure algorithmic or structural similarity between a pair of source codes and are accurate in finding similar source codes. However, two important issues are yet to be discussed. Firstly, the similarity is not always caused by plagiarism; it could be caused by coincidence. Secondly, instructors might feel the burden of asking students if they have plagiarized someone's source code, and students might feel the burden of being asked. In this paper, we introduce our novel index for source code plagiarism detection, which can be utilized to check the cause of the similarity outputted from existing methods, report the result of questionnaire investigation with 89 freshmen in our junior college, and discuss the possibility of reducing psychological burdens in source code plagiarism detection with our method.
ISSN:0190-5848
2377-634X
DOI:10.1109/FIE.2010.5673650