Test case minimization and prioritization for regression testing using SBLA-based adaboost convolutional neural network

The software engineers retain the test cases they create for specific software for future usage. This form of test case reuse is called regression testing and this step mainly improves the software testing efficiency. The test case minimization and prioritization for regression testing raise different issues such as higher time consumption and heavier resource utilization. To overcome this problem, this paper presents a Side-blotched lizard optimized AdaBoost Convolutional Neural Network (SBLA-AdaBoost CNN) model. The proposed technique mainly aims to discover the faults initially and minimize the test case execution cost. The proposed model is evaluated using the Defects4J dataset. Our proposed method tends to be cost-effective since it integrates test case selection, prioritization, and minimization. The proposed methodology can be also utilized to arrange the test cases during their initial stages of software testing. The results demonstrate that the proposed methodology is efficient in identifying the changes in different parts of the source code, minimizing resource utilization, and time consumption. The precision and recall score obtained by the proposed methodology is 98.5% and 99% which is relatively higher than the state-of-art techniques. The time taken by the proposed methodology to evaluate a total of 50 test cases is 19.14 s.

Automated summary

This paper proposes a model based on Side-blotched lizard optimized AdaBoost Convolutional Neural Network (SBLA-AdaBoost CNN) to minimize test case execution cost and improve efficiency. The model is evaluated using the Defects4J dataset and shows high precision and recall scores, as well as reduced resource utilization and time consumption. This method can be applied during the initial stages of software testing.