Dynamic mutation testing in integrated regression analysis

A new method of integrated regression analysis is proposed. Its core is the clustering, a method for automatic identification of program modifications. Clustering is used to formulate a hypothesis about the existence of a fault in the modified program, and to guide the process of testing this hypothesis. It is postulated that static and dynamic program analysis be used for that purpose. Specifically, the authors introduce dynamic mutation testing (DMT), an experimental technique for testing the fault hypothesis. DMT estimates the sensitivity of the test-induced program state to reveal the postulated fault in the modified program. If all tests pass and are found to have a high sensitivity, the fault hypothesis can be rejected at a high level of confidence