Clustering test cases to achieve effective test selection

Abstract

Full testing involves running all the tests in the test suite. This is exhaustive and will consume an inordinate amount of time and money. Hence, an ordering of test cases aids in early detection of faults. However, ordering and running a large test suite is still infeasible, as it would not be possible to run all tests during regression testing. In this work, clustering is used to select a subset of scenarios for testing. First, a distance matrix is obtained by using Levenshtein distance to compare scenarios. This distance matrix is used as input for the Agglomerative Hierarchical Clustering(AHC) technique with the objective of selecting dissimilar test scenarios and at the same time achieveing maximum coverage and rate of fault detection. © 2010 ACM.