Test prioritization at different modeling levels

Abstract

Validation of real-life software systems often leads to a large number of tests; which, due to time and cost constraints, cannot exhaustively be run. Therefore, it is essential to prioritize the test cases in accordance with their importance the tester perceives. This paper introduces a model-based approach for ranking tests according to their preference degrees, which are determined indirectly, through event classification. For construction of event groups, Gustafson - Kessel clustering algorithms are used. Prioritizing is performed at different granularity levels in order to justify the effectiveness of the clustering algorithm used. A case study demonstrates and validates the approach. © 2010 Springer-Verlag Berlin Heidelberg.