Machine Learning Approach in Mutation Testing

Abstract

This paper deals with an approach based on the similarity of mutants. This similarity is used to reduce the number of mutants to be executed. In order to calculate such a similarity among mutants their structure is used. Each mutant is converted into a hierarchical graph, which represents the program's flow, variables and conditions. On the basis of this graph form a special graph kernel is defined to calculate similarity among programs. It is then used to predict whether a given test would detect a mutant or not. The prediction is carried out with the help of a classification algorithm. This approach should help to lower the number of mutants which have to be executed. An experimental validation of this approach is also presented in this paper. An example of a program used in experiments is described and the results obtained, especially classification errors, are presented.

Applying Mutation Testing for Assessing Test Suites Quality at Model Level

“…It seems that the migration to the model level alone lowers the number of generated and executed mutants, but other techniques should also be considered. The most efficient techniques should be the ones taking into account individual characteristics of modeled systems, such as proposed in [20], [21], [22].…”

Using Classification for Cost Reduction of Applying Mutation Testing

“…Previously graph edit distance and a distance computed from simple HCFG kernel were used in this algorithm [4]. In this paper distances are computed from the three new kernels described in sections IV-C and IV-D. (Denoted as Ker2HCFG -defined by equation 4 and Map1HCFG and Map2HCFG -by equations 6 and 7, respectively).…”

“…The kernel defined by equation (4) is a modification of the one proposed in [4], [41], as it uses a binary node kernel given by definition 2 to compare the neighbouring nodes rather than the more complex k V kernel which was previously used. It lowers the number of times a tree kernel is computed and does not affect the accuracy of the classification.…”

“…In this paper we propose to use a so called hierarchical control flow graph (HCFG), which is a combination of CFD and hierarchical graphs [40], [4]. Such a graph adds a level of hierarchy to the traditional control flow diagram.…”

“…The paper presents a continuation of an ongoing research on using methods of machine learning to reduce the costs of applying mutation testing [4], [5]. Mutation testing [6] is a recognized software testing technique used to support selection of tests [7], [8].…”