To survive in competitive marketplaces, most organizations have adopted agile methodologies to facilitate continuous integration and faster application delivery and rely on regression testing during application development to validate the quality and reliability of the software after changes have been made. Consequently, for large projects with cost and time constraints, it is extremely difficult to determine which test cases to run at the end of each release. In this paper, a test case prioritization and selection approach is proposed to improve the quality of releases. From existing literature, we analyzed prevailing problems and proposed solution relevant to regression testing in agile practices. The proposed approach is based on two phases. First, test cases are prioritized by clustering those test cases that frequently change. In case of a tie, test cases are prioritized based on their respective failure frequencies and coverage criteria. Second, test cases with a higher frequency of failure or coverage criteria are selected. The proposed technique was validated by an empirical study on three industrial subjects. The results show that the method successfully selects an optimal test suite and increases the fault detection rate (i.e., more than 90% in the case of proposed technique and less than 50% in other techniques), which reduces the number of irrelevant test cases and avoids detecting duplicate faults. The results of evaluation metrics illustrate that the proposed technique significantly outperform (i.e., between 91 and 97%) as compared to other existing regression testing techniques (i.e., between 52 and 68%). Therefore, our model enhances the test case prioritization and selection with the ability for earlier and high fault detection. Thus, pruning out irrelevant test cases and redundant faults and enhancing the regression testing process for agile applications.

中文翻译：

---

用于敏捷软件开发和持续集成策略的增强型回归测试技术

为了在竞争激烈的市场中生存，大多数组织采用了敏捷方法来促进持续集成和更快的应用程序交付，并在应用程序开发期间依靠回归测试来验证更改后软件的质量和可靠性。因此，对于具有成本和时间限制的大型项目，确定在每个版本结束时运行哪些测试用例是极其困难的。在本文中，提出了一种测试用例优先级和选择方法来提高发布质量。从现有文献中，我们分析了与敏捷实践中的回归测试相关的普遍问题并提出了解决方案。建议的方法基于两个阶段。首先，通过将那些经常更改的测试用例聚类来确定测试用例的优先级。在平局的情况下，测试用例将根据它们各自的失败频率和覆盖标准进行优先级排序。其次，选择具有较高失败频率或覆盖标准的测试用例。所提出的技术通过对三个工业主题的实证研究得到验证。结果表明，该方法成功地选择了一个最优的测试集，提高了故障检测率（即提出技术的情况下超过90%，其他技术的情况下不到50%），从而减少了无关测试用例的数量和避免检测重复故障。评估指标的结果表明，与其他现有回归测试技术（即，介于 52 和 68％之间）相比，所提出的技术显着优于（即，介于 91 和 97％之间）。所以，我们的模型通过早期和高故障检测的能力增强了测试用例的优先级和选择。因此，修剪不相关的测试用例和冗余故障并增强敏捷应用程序的回归测试过程。