Formation of Weighting Coefficients for Selection Methods for Regression Testing

The article discusses methods for ranking and grouping tests to form test suites during regression testing: by priority, execution duration, previous result, total number of detected errors, deadline, microservice priority, volume of changes made to the microservice, and the presence of modifications in related services. An approach is proposed for generating relevant weighting coefficients for each test selection and ranking method, resulting in a resulting test suite, so that tests capable of detecting the most important errors in an information system are located in the top half of the list. Application methods for generating weighting coefficient values for two test management systems are described. An example of using the proposed approach to forming a ranked list of tests for regression testing of a real microservice information system is considered using an automated test ranking system, which allows any number of ranking methods to be connected to modern test management systems. Graphs are provided for depreciation of weighting coefficients and the appearance of missed defects depending on the completeness of the test suite execution.