Homologous detection based on text, Token and abstract syntax tree comparison

Abstract

The research of software homologous detection is more important under the situation that the software market flourishes. Most of current detection tools are based on text which ignores the syntax structure of program codes, a few tools based on syntax structure are not good enough. CodeCompare the homologous detection system in this paper combines compare methods of text, Token, and abstract syntax tree using improved LCS algorithm as string comparison algorithm, which can detect the homologous of software more accurately and comprehensively. The comparison based on abstract syntax tree considers and analyses the syntax structure of program codes, achieves the detection on syntax aspect. Compared to other detect tools, CodeCompare includes more code plagiarizer methods, can detect more effectively.