WASTK: A Weighted Abstract Syntax Tree Kernel Method for Source Code Plagiarism Detection

Abstract

In this paper, we introduce a source code plagiarism detection method, named WASTK (Weighted Abstract Syntax Tree Kernel), for computer science education. Different from other plagiarism detection methods, WASTK takes some aspects other than the similarity between programs into account. WASTK firstly transfers the source code of a program to an abstract syntax tree and then gets the similarity by calculating the tree kernel of two abstract syntax trees. To avoid misjudgment caused by trivial code snippets or frameworks given by instructors, an idea similar to TF-IDF (Term Frequency-Inverse Document Frequency) in the field of information retrieval is applied. Each node in an abstract syntax tree is assigned a weight by TF-IDF. WASTK is evaluated on different datasets and, as a result, performs much better than other popular methods like Sim and JPlag.

TF-IDF Inspired Detection for Cross-Language Source Code Plagiarism and Collusion

“…These techniques rely on various structures for comparison. Some of the structures are source code token strings [8,29,41], abstract syntax trees [20,30,53], parse trees [48], program dependency graphs [32], and low-level token strings [25,42].…”

Automated Hints Generation for Investigating Source Code Plagiarism and Identifying The Culprits on In-Class Individual Programming Assessment

“…Source code token sequence [19][20][21] is a sequence of meaningful "words" from source code; it is usually extracted with the help of a programming language parser. Abstract syntax tree [22] is a tree where the nodes are formed from tokens and the edges are formed from programming language grammar. Program dependency graph [23] is a graph connecting several instructions based on their execution dependencies.…”

A Language-Independent Library for Observing Source Code Plagiarism

“…Structure-based tool defines source code similarity based on given codes' shared structure. The structure can be either source code token sequence [2], [15], [16], low-level token sequence [17], [18], program dependency graph [19] or abstract syntax tree [20]. The similarity of the first two structures are commonly measured by string matching algorithms (e.g., Running-Karp-Rabin Greedy-String-Tiling [5]), that have been modified to handle source code tokens instead of characters.…”