基于写作风格特征的论文剽窃检查优化方法研究

发布时间：2018-05-05 23:08

本文选题：写作风格特征 + 票窃检查　；参考：《复旦大学》2011年硕士论文

【摘要】：互联网技术日新月异的发展与网络数据库资源的日益丰富,为科研工作带来极大的帮助。学术论文、调研报告、分析数据等等学术论文写作所需要的参考资料得以便捷获取,与此同时论文抄袭也相应地更为容易与常见。寻找并建立有效预防及遏制剽窃行为的手段已经刻不容缓。自2005年以来,作者所在课题组通过产学研合作模式,在论文剽窃检查方面进行了大量的研究和开发,完成了基于词频的论文剽窃检查以及基于相对单元密度的论文剽窃检查的设计与实现。前者对于完全抄袭的情况可以起到很好的判别作用,后者则在此基础上完成了对部分抄袭情况的判断,使得检查结果的召回率得到显著提高。然而,这两种剽窃检查方法在改变原文的剽窃行为判断方面还存在较大的局限。为此,我们在其基础上引入了综合性考量对象——写作风格特征,对现有的剽窃检查方法进行优化。主要工作有如下4个方面： 1.本文研究对比了国内外主流的与写作风格特征分析相关的技术以及语义词典技术,从中寻找适合应用于单篇论文的,满足剽窃检查应用需求的技术思路。 2.介绍了本课题组的前期工作：设计并实现了基于词频统计的论文剽窃检查算法,以及基于相对单元密度的论文剽窃检查应用。在介绍前期工作取得的具体进展同时,还说明了目前这两个方法存在的问题、局限以及可改进之处。 3.在前期工作基础上,借鉴国内外相关技术,提出了基于写作风格特征的论文剽窃检查优化方法,建立初步的写作风格特征语义词典,描述了相应的论文剽窃检查系统的结构与整体流程。 4.本文通过具体的应用实例分析,阐述了优化方法的应用场景与效果,验证了新方法的有效性。本文所研究的基于写作风格特征的论文剽窃检查方法是对前期工作的补充优化,对改变原文的论文剽窃情况进行分析检查,为剽窃检查课题引入了新的思路,帮助该课题进一步深入研究奠定基础,从而逐步建立起更准确更完善的剽窃检查方法与系统,对学术剽窃的不正风气起到有效的打击预防作用。
[Abstract]:The rapid development of Internet technology and the increasing abundance of network database resources bring great help to scientific research. Academic papers, research reports, data analysis and other academic papers required for the writing of reference materials can be easily obtained, at the same time, the paper plagiarism is also easier and more common. It is urgent to find and establish effective means to prevent and curb plagiarism. Since 2005, the author's research group has carried out a lot of research and development in the area of plagiarism inspection through the cooperation model of industry, education and research. The thesis plagiarism check based on word frequency and the paper plagiarism check based on relative unit density are designed and implemented. The former can play a very good role in discriminating the situation of complete plagiarism, while the latter has completed the judgment of partial plagiarism on this basis, which makes the recall rate of inspection results improved significantly. However, these two methods of checking plagiarism still have some limitations in changing the judgment of plagiarism. On the basis of this, we introduce the comprehensive object-writing style feature to optimize the existing methods of checking plagiarism. The main tasks are as follows: 1. In this paper, the main technologies related to the analysis of writing style and semantic dictionary are compared, and the technical ideas suitable for the application of plagiarism inspection are found. 2. This paper introduces the previous work of our group: we design and implement the algorithm of checking plagiarism based on word frequency statistics and the application of checking plagiarism based on relative unit density. At the same time, the problems, limitations and improvements of these two methods are explained. 3. On the basis of previous work and drawing lessons from relevant technologies at home and abroad, this paper puts forward an optimized method of checking plagiarism based on writing style features, and establishes a preliminary semantic dictionary of writing style features. The structure and overall flow of the corresponding paper plagiarism checking system are described. 4. In this paper, the application scene and effect of the optimization method are expounded through the analysis of the concrete application examples, and the validity of the new method is verified. The method of checking plagiarism based on writing style in this paper is a supplementary optimization to the previous work. It analyzes and checks the plagiarism situation of the original text, and introduces a new way of thinking for the subject of plagiarism checking. In order to establish a more accurate and perfect method and system for checking plagiarism, it can effectively combat and prevent the abnormal trend of academic plagiarism.
【学位授予单位】：复旦大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：TP391.1

【参考文献】