TEST CASE PRIORITIZATION BASED ON REQUIREMENTS CORRELATION AND  SEVERITY OF FAULTS

DANIEL, ADDIS

dc.contributor.author	DANIEL, ADDIS
dc.date.accessioned	2022-11-17T11:44:34Z
dc.date.available	2022-11-17T11:44:34Z
dc.date.issued	2022-07
dc.identifier.uri	http://ir.bdu.edu.et/handle/123456789/14444
dc.description.abstract	Software development has several stages to be followed by developers, starting from planning to deployment and maintenance. One of the most important steps in the software development life cycle is software testing. Software testing is a process used to check the correctness, completeness, and the quality of developed software. Regression testing is a type of testing which encompasses the activities of retesting a software system after changes have occurred. Due to limited resources, just a subset of test cases is usually executed for each release. As a result, identifying test cases which are most likely to detect the majority of errors becomes a challenge. To generate efficient test cases in regression and other testing approaches, test case prioritization has paramount advantages. Taste case prioritization can be done based on different criteria. In this research we have employed requirement correlation and fault severity to build a model for test case prioritization. Previous researchers designed test case prioritization techniques based on requirements. But they applied hard computing algorithms, which implies inflexibility and imprecision of results. To perform this study, we used an experimental research approach. For the experiment, we have prepared one thousand (1000) test cases which were labeled by experts as positive and negative. The datasets that were fed to our proposed model were pre-processed using Natural Language Processing (NLP) principles. We used Term Frequency-Inverse Document Frequency (TF-IDF) vectorization after preprocessing to convert the textual data format to vectorized form. The machine learning techniques applied to build the model are Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Naïve Bayes (NB), and Decision Tree (DT). To reach and identify a model that had high performance, we executed twelve different combinations of experiments. The accuracy obtained from each experiment ranges from 75% to 94%. SVM based on requirement correlation and severity of fault outperformed for test case prioritization. When we consider the severity of a fault, only KNN is the best technique. The reason behind this result is that KNN is good for a small number of features, and SVM is a better technique for a large number of features. Keywords: Machine Learning, NLP, Software testing. Test Case, Test Case Prioritization	en_US
dc.language.iso	en_US	en_US
dc.subject	FACULTY OF COMPUTING	en_US
dc.title	TEST CASE PRIORITIZATION BASED ON REQUIREMENTS CORRELATION AND SEVERITY OF FAULTS	en_US
dc.type	Thesis	en_US