Mining usage profile to improve web performance testing

Abstract

Testing is an important means to ensure the high-quality, high-performance and high-reliability of web application system. This paper focuses on how to improve the testing performance by using usage profile generation method. First, apply the web mining technology to performance test, proposes usage profile generation method based on historical data mining, include the log file preprocessing, presents a frequent access pattern tree generation algorithm to dig out user's frequent access pattern, analyses user behaviour characteristics, and constructs user usage profile approximate to the real usage. Then applied the usage profile into the performance test design stage, and guide test scenarios; Finally, by comparison experiments to verify that the proposed usage profile can effectively guide the process of test design, and can improve the accuracy of performance test results.