Research On Techniques For Reducing User Session Data Sets Base On Concept Lattice In Web Testing

User-session based testing technologies mainly rely on the collection of user session data sets, the more data, testing efficacity the better. However, with increasing quantity of data, the cost of analysis and storage these user session will rapidly increase. As a result, there is need to find a balance between collect data the scale of collecting data testing effectiveness.This dissertation presents a new technique for reduceing user session based on the concept lattice. First, the comparison of several traditional batch concept lattice algorithms and design a more efficient batch concept lattice algorithms. Using this algorithm to clustering the initial user session data sets , generates the concept lattice. then, making use of the all-node, next-to-bottom-node or k-level nodes choosing strategies, users can choose user session data from concept lattice to buildup a test case subset instead of the initial data set for testing. Furthermore, when a new user session added, using the incremental concept lattice algorithm to analyze additional user sessions, generate an updated concept lattice, and re-select the user session to compose a new subset, with respect to incrementally update the reduced test suite. Finally, designed a test tools, to achieve automatic collection, reducing, updating user session data sets, replay test case and the generation of test reports.The new technology applies to large-scale user session data sets, which can be selected user session in the initial data sets and buildup a smaller subset .Useing this subset to test, can decrease the testing cost, while still guarantee defect detection, program coverage, and another testing effectiveness to the same with the testing requirements of initial set. The new technology can effectually resolve the problem between the quantity of data and testing effectiveness in user-session based testing.