學生程式碼相似度之研究—以抄襲偵測之應用為例

Research on Detection of Similarity in Student Programs – with Application to Detection of Plagiarism

林世唐(Shih-Tang Lin)

指導教授 ： 魏世杰

淡江大學/管理學院/資訊管理學系碩士班/碩士(2005年)

https://doi.org/10.6846/TKU.2005.00513

摘要

程式的相似判斷不像文字那麼複雜，比起文字的語法來說，程式的文法更有
規則性。程式相似判斷的應用上有很多方面，在實際教學應用上，最常用來作抄
襲的檢測。但我們發現在可以參考範例的考試或作業下，可能因為修改自相同的
範例程式，因此雖然找出大量相似片段組，但有很多相似是因為參考範例本身而
相似，不具抄襲意義。另一方面，在資訊檢索領域有IDF (Inverse Document
Frequency, 反轉文件頻率法)的概念，發生頻率高的片段較不具意義，發生頻率
低的片段較具意義。因此我們提出以IDF 為主的新方法，幫助我們找出發生頻率
低的相似片段組，視為較有抄襲可能性之片段。並用開放式(open book)的一次考
試和一次作業程式來作驗證。

關鍵字

程式比對 ； 結構度量 ； 抄襲偵測 ； 反轉文件頻率法

並列摘要

Similarity detection on programs is simpler than on text documents. Compared to
text documents, the grammar used in program languages is easier to define. As a
result, more and more applications are developed to detect program similarity. One
practical use of these appications is to detect plagiarism for educational purposes. In
particular, when students have test or homework on programs where they can open
books to consult the examples, they may copy from the example programs without
much thinking and rewriting. In this case, we will find many similar code tiles that are
copied from the same example but of little value in plagiarism detection. So in this
paper based on the IDF (Inverse Document Frequency) concept from information
retrieval ,we propose a new method to reduce the influence of high frequency code
tiles, and compare with traditional non-IDF result using the datasets from an
open-book program test and a homework.

並列關鍵字

Program Compare ； Structure Metric ； Plagiarism Detection ； Inverse Document Frequency

參考文獻

1. P.Clough, “Plagiarism in Natural and Programming Languages: an Overview of

of Sheffield, July 2000.

Plagiarism Detection and Its Practical Implementation”, University of California,

Santa Barbara, May 5, 2002.

3. J.L. Donaldson, A.M. Lancaster and P.H. Sposato,” A Plagiarism Detection