基于关键词相关度的计算机辅助定密技术研究

发布时间：2018-07-17 15:56

【摘要】：随着计算机信息化技术的飞速发展与普及,信息化技术在越来越多的的领域得到应用与推广。保密技术领域当然也不例外,但随着应用管理的深入,问题同时也显现出来,其中之一就是在信息化条件下文档密级确定上存在的效率不高、准确率有待提升、经验传承可持续性不强等问题。目前很多单位定密工作都采用人工定密的方式,定密流程不规范且职责不明确,定密人的权限划分不明确,甚至还存在"拍脑袋定密"的问题。针对上述问题,本文提出了一个基于关键词相关度分析的定密方法,并在此基础上设计了计算机辅助定密系统,既能提高定密工作准确性,又能很大程度提高工作效率。。本文首先对定密工作的相关背景进行介绍,在对已有技术进行介绍的基础上,将基于词激活力的文档语义网络应用于文档向量表示模型中,并在此基础上计算文档间的相似度,提出了一种利用分类思想为待定密文档确定密级方案。首先,论文研究了基于领域关键词对文档进行分词并进行向量化表示方法;其次,基于词激活力构建文档的语义网络,为文档的向量化表示模型增加语义层特征;再次,以文档的增强型向量表示及基础,计算待定密文档与已定密样本文档的相似性;最后基于文本相似度确定待定密文档的密级。最后,基于上述定密方法,本文设计并实现了一个计算机辅助定密系统软件,并对系统性能进行验证。测试证明,本文提出的方法能够有效辅助定密人员对目标文档进行快速标密,具有较高的准确率、并且消耗时间较短,具有很高的实用价值。
[Abstract]:With the rapid development and popularization of computer information technology, information technology has been applied and popularized in more and more fields. Of course, the field of secret technology is no exception, but with the deepening of application management, the problems also appear. One of the problems is that the efficiency of document classification is not high and the accuracy rate needs to be improved under the condition of information technology. The sustainability of memory transmission is not strong and so on. At present, many units adopt the method of manual secret, and the process of keeping secret is not standard and the duty is not clear, the authority of the secret person is not clear, and even there is the problem of "beat the head and fix the secret". In order to solve the above problems, this paper presents a keying method based on keyword correlation analysis, and then designs a computer-aided encryption system, which can not only improve the accuracy of classified work, but also greatly improve the working efficiency. This paper first introduces the relevant background of the work of encryption. Based on the introduction of existing techniques, the semantic network of document based on word activation force is applied to the document vector representation model, and the similarity between documents is calculated on this basis. In this paper, a scheme of classifying secret documents to be determined by classification is proposed. Firstly, this paper studies the segmentation and vectorization of documents based on domain keywords. Secondly, the semantic network of documents is constructed based on word activation force, which adds semantic layer features to the vectorization model of documents. Based on the enhanced vector representation and the basis of the document, the similarity between the undetermined secret document and the pre-encrypted sample document is calculated. Finally, the secret level of the undetermined secret document is determined based on the text similarity. Finally, based on the above encryption method, a computer aided encryption system software is designed and implemented, and the performance of the system is verified. The test results show that the proposed method can effectively assist the steganographer to mark the target document quickly, has a higher accuracy, and consumes less time, so it has a high practical value.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：D631.31;TP391.7

【参考文献】