A Method for Test Cases Reduction in Web Application Testing Based on User Session

Abstract

With the development of web application, the internal structure of the web application was more complicated. The software testing method based on user session has been used to verify the quality of web application. However, the amount of real user session data were extremely large. In order to reduce the scale of the test case set for web application testing, we proposed a reduced method of test cases based on user session. First, we cleaned the original log file and identified user session and sorted the user sessions according to the Gini index. Then, we made an “OR” operation on the all user sessions coverage to obtain some user sessions that cover all pages of web application. These user sessions were a set of reduced test cases. Finally, the structure dependency graph of web application was generated according to the user sessions. The PageRank algorithm and the Hamming distance are used to optimize the test cases reduced with the same Gini index. By this means, we obtained the final test cases set for reduction. Through experimental verification, we found that the reduction rate of the test cases reached 85.7% and the fault coverage rate reached 100% by our method. Therefore, the method we proposed in this paper improved effectively the reliability of web application and the effectiveness of software testing and reduced the cost of software testing.