文章基本信息

标题：IR-based technique for linearizing abstract method invocation in plagiarism-suspected source code pair
本地全文：下载
作者：Oscar Karnalim
期刊名称：Journal of King Saud University @?C Computer and Information Sciences
印刷版ISSN：1319-1578
出版年度：2019
卷号：31
期号：3
页码：327-334
DOI：10.1016/j.jksuci.2018.01.012
出版社：Elsevier
摘要：According to several works, low-level approach is an effective and efficient solution for detecting source code plagiarism. Instead of relying on source code tokens, it compares the executable form of given code; that form only contains semantic-preserving tokens and is resistant to various plagiarism attacks. However, to our knowledge, an issue about statically linearizing abstract method (i.e. replacing each abstract method invocation with its respective invoked method content without considering invocation semantic) has not been handled comprehensively. Such issue, at some extent, will generate inaccurate plagiarism detection result when handling object-oriented source codes. This paper aims to solve such issue locally per plagiarism-suspected pair. It will generate all possible linearization pair alternatives and select the correct one through IR-based similarity. According to our evaluation regarding the reversed number of mismatched token, the number of false positive, and the number of process, proposed technique is more effective and efficient when compared to state-of-the-art and combinatoric technique. In addition, it is also observed that each IR mechanism used in proposed technique has its own exclusive benefit for selecting the correct linearized forms.
关键词：Abstract method linearization ; Source code plagiarism ; Software engineering ; Information retrieval