To investigate the correctness of different approaches to regression testing, a model for estimating regression test methods is used. The main concepts to consider are completeness, accuracy and efficiency.
Completeness is a criterion measure for selection of tests from T set, in which the result of executing a changed program is different from the result of the initial program, therefore errors can be detected in P’ set. The 100% complete method is called a safe method.
Accuracy is a measure of the ability of a method to avoid selecting tests from T set in which the result of executing a modified program will not differ from the result of its original version, namely, tests that are unable to detect errors in P ‘.
Quality assurance consultant provides software designers with useful information about how to ensure success of a development project. This person advises the clients on the most effective approaches to building marketable IT products.
Based on the above definition, the accuracy of a test set is the ratio of the number of tests of a given set in which the results of the execution of the new and old programs differ, to the total number of tests in the set. Precision is an important attribute of regression testing. An inaccurate method tends to select tests that should not have been selected. The less accurate the method, the closer the volume of the selected test suite approaches the volume of the initial set of tests.
Efficiency is the estimation of the computational cost of a selective regression testing strategy, specifically, the cost of implementing its requirements in respect to time and memory, as well as the automation capabilities. Relative efficiency is the comparison between a given procedure and the best possible procedure. Absolute efficiency is the ability of a test method to work effectively in real-world conditions while it does not matter how many errors are present in the program being analyzed.