Researches And Implements Of Text Filtering For Physical GAP

With the Internet and E-Government Network being widely used in working area, life as well as study area, network security is being concerned increasingly, security of network information is focused while developing the network in whole society, invention and development of Physical Gap based on the isolation technology is in accordance with the requirement of network security. Text filtering being applied currently for Physical Gap mostly utilizes key words filtering and URL filtering which has poor performance on precession and efficiency for filtering and which also are not able to meet the today's requirement for network security.With research and development for text filtering technology, especially the area for text representation with VSM and filtering algorithm based on VSM is becoming more and more mature, it is helpful for us to make out a advanced text filtering technology to replace key words filtering and URL filtering currently being used for Physical Gap in order to improve the text filtering performance for Physical Gap and also meet more stricter requirement of network security. Research and implement for Intelligent Text Filtering solution for SGAP is a main task of this paper.Author proposes one solution of intelligent text filtering by using some relevant techniques such as filtering algorithm model, information processing, rough set theory and physical gap, the solution is based on statistics and suitable for physical gap to improve current key words filtering in SGAP. A detailed introduction for critical technology in new solution such as hybrid feature selection and modified KNN algorithm is made.Hybrid feature selection is to combine traditional one of feature selection methods with rough set attribute reduct method, primarily use one of feature selection methods to select features, followed by further select features using rough set attribute reduct and more accurate and few features are extracted. Modified KNN algorithm is based on traditional KNN algorithm and text cluster theory and not only helpful for filtering effectiveness but also filtering speed. Comparing with traditional KNN algorithm, modified KNN algorithm is a right solution and more suitable for text filtering solution in physical gap application.