Research Onthe Computer-aided Secret-level Classification System Based On Keywords Relevancy

With the rapid development of computer information technology and popularization,the information technology has been applied and spread in more and more fields included confidential technology.But with the deepening of the application management,problems are exposed at the same time.Based on the condition of informatization,one of the problems is the classification of secret-level,which should be improved in efficiency,accuracy,inheritance and sustainability.At present,many units classify the secret-level by artificial method.The process is not standard and the division of responsibility and authority is not clear.Even some secret-level is defined by " smacking forehead ".Focusing on these problems,the thesis present a method based on analysis of keywords relevancy.Based on the above method,designed computer aided system which can greatly improve the accuracy and efficiency on classifying document secret-level.The thesis introduces the background of classification of secret-level.On the basis of introduction about existing technologies,we calculate the text similarity by using the dynamic semantic network which is based onword activation forces,and then,determine the classification of the document.Firstly,the thesis studies themethod of words segmentation and vectorial representationof document based onkeywords;secondly,we build dynamic semantic network by using the word excitation energy,to increase the semantic layer characteristics for document quantitative models.Thirdly,based on the enhanced vectors of document,we calculate the similarity of undetermined secret-level document with determined secret-level document.Finally,on the basis of the text similarity,we classify the secret-level of undetermined secret-level document.Finally,based on the above method of classification of secret-level,we design and realize a computer-aided secret-level classification system,whose performance was verified.Test results have shown that,the method can assist staff to define secret-level of the target documents.From what has been discussed above,the method has high accuracy and efficiency,which make it has outstanding practical value.