基于用户分类的web日志统计测试研究  

Research of web log statistical testing based on user classification

作　　者：俞金松[1] 高建华[1] 

机构地区：[1]上海师范大学计算机科学与工程系,上海200234

出　　处：《计算机工程与设计》2012年第12期4551-4556,共6页Computer Engineering and Design

基　　金：国家自然科学基金项目(61073163);上海市教委科研创新重点基金项目(0922135);上海市科委基金项目(09220503000)

摘　　要：针对传统统计测试方法效率低和适应性不强的局限性,结合web应用的特点,提出了一种基于用户分类的web日志统计测试方法。根据web应用的不同复杂度,通过用户分类和web日志统计进行建模,并依据该模型测试,评估web应用的可靠性。实验结果表明,该方法较传统统计测试方法对于web应用主要业务功能测试的覆盖率更高,其可靠性评估更具现实意义。To solve the limitations of low efficiency and inflexible for traditional statistical testing, a method of Web log statistical testing based on user classification with the characteristic of Web applications is proposed. According to the difference complex of Web applications, the user groups are classified and the Web information is extracted from Web usage and failure logs. Using this information, models for Web log statistical testing is built and the reliability of Web application is estimated. An actual Web application is applied by this approach. The results demonstrate this approach can support higher testing coverage of main func- tions for Web application than the traditional statistical testing does, and the reliability assessment are more viable and realistic.

关 键 词：统计测试 WEB日志 马尔可夫链模型 用户分类 可靠性分析 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]