Research On Technologies Of User Access Behavior Check Based On Web Data Stream Analysis

With the rapid development of the Internet, information systems are increasingly beingapplied to all walks of life to play an important role for the improvement of production efficiency,to facilitate people’s daily lives. At the same time, the security issues as the constraints networkkey to healthy development factors, many of which are due to improper user or user breachescaused by malicious access caused. Check of the user’s access behavior, has become one of thefocus of research in the field of information systems security.To solve the above problem, the data stream to Web information system as the researchobject, Check network access behavior through data flow analysis. The main tasks include thefollowing:1.Check requirements from user access Web-based data flow analysis, the analysis of Webtraffic generation process and Web information system user access process, summed up the Webtraffic bi-directional, dynamic, correlation characteristics. Information system applicationenvironment is characterized by raised and the establishment of an adaptation to the Webapplication environment based on data flow analysis, user access verification framework, theframework is transparent to the application system to accurately track user access and efficientverification of user behavior features information system users transparent access behavior,effective verification laid the foundation.2. Web user based on the properties of the multi-dimensional identity identification andtracking method based on the above framework. For the network data stream the user identitydebris, dynamic, diversity, user identification and tracking method based on multi-dimensionaluser identity attribute. Firstly, the identity of the user property is divided into specific attributesand dynamic attributes describing the attributes into three categories, three-dimensionaldescription of the user’s identity; Then, based on data stream user identity fragment analysis,design of user identity fragment identification strategies; Last user identity-based fragmentassociated user identification algorithm based on identity attributes associated with user accesstracking algorithm. The method has high identification accuracy and integrity and good trackingof user access behavior.3、Gives the user access verification methods to support the regression error correction.Existing based policy network behavior verification methods, high efficiency, but does notsupport dynamic changes in user behavior, does not support the verification of the unknownbehavior patterns. Existing verification methods based on indicators of network behavior,applicability, but the efficiency is relatively low. In this paper, first proposed to supportregression error correction user access verification program: known user access to network behavior based on the description method designed on the basis of the recognition algorithmbased on network behavior; unknown user access, user behavior evaluation method based onanalytic hierarchy process, optimizing behavior verification policy repository based on theevaluation results. The method of verification through learning, feedback mechanisms, adjustand optimize the behavior policy repository and verification indicators, the method supports bothknown behavior and unknown behavior verification experiments show that it has a highverification accuracy and better adaptability.4. On the basis of in-depth study of the technology, user access verification prototypesystem based on Web data flow analysis on some of the features to achieve authentication, andrelated projects.