基于篇章结构相似度的复制检测算法

学术论文的复制检测研究对于知识产权保护和抑制抄袭侵权等行为有重要意义.国内外主要用数字指纹及关键词匹配等技术进行论文的复制检测.为解决目前中文复制检测难题,给出了一种基于篇章结构相似度的中文学术论文复制检测算法及其问题的数学模型.在分析论文篇章结构的基础上,利用数字指纹和词频统计等技术,经编程实现,用于论文的全抄、部分抄袭和拼抄等抄袭现象的初步检测.与基于全文数字指纹和基于全文词频统计的检测方法相比较,更适用于要求较准确的论文复制检测.

Document-structure-based copy detection algorithm

Research on copy detection of academic papers is important in both intellectual property protection and academic plagiarism prevention.Nowadays,researchers mainly use digital fingerprinting technique and keyphrase matching technique on copy detection.To overcome the difficulty of Chinese copy detection,a set of document-structure-based algorithm for identifying Chinese plagiarized papers is presented,and mathematical models on it are given.The plagiarism identification(including full-plagiarism,part-plagiarism and pieced-plagiarism) is realized with the help of document-structure analysis,fingerprinting and word-frequency techniques.Lastly,comparing with two typical identification methods,the effectiveness of accurate paper copy detection of the proposed algorithm is demonstrated.