• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Journal of computer sciences >Comparing Approaches for Quality Evaluation of Software Engineering Experiments: An Empirical Study on Software Product Line Experiments
【24h】

Comparing Approaches for Quality Evaluation of Software Engineering Experiments: An Empirical Study on Software Product Line Experiments

机译：软件工程实验质量评估的比较方法：软件产品线实验的实证研究

获取原文
获取原文并翻译 | 示例
           
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The Software Engineering (SE) research area must provide results of a certain quality for the sake of value. High quality research results may ensure experience and knowledge, which are essential for the technology to be transferred to the industry. One of the means to obtain such quality results is experimentation. Experimentation is a scientific method that aims to provide evidence of a theory over real-world observations establishing a cause-effect relation. Well conducted, auditable and repeatable experiments are vital for scientific evolution and novelty. Quality evaluation of controlled experiments and quasi-experiments in SE has been recently discussed in the literature as researchers desire to assess whether such experiments have improved by reporting information that enables the experiments to be replicated and the reader can understand the experiment and validate results. Thus, this work empirically compares four approaches for quality evaluation of SE experiments in the context of Software Product Lines (SPL). In addition, we are interested on verifying the quality of reporting experiments in a well-discussed reuse technique as SPL. The Pearson technique supported the correlation between pairs of evaluation approaches. In addition, the T-Test and Mann-Whitney-Wilcoxon U test were applied to the samples to verify whether there was a difference in the quality of experiments when using an experimental template. Preliminary results show a strong positive correlation between them, the hypothesis tests confirmed there is such a difference in quality when using experimental template and the SPL experiments report more the planning phase than the analysis and interpretation phase. Based on our results, we provide initial evidence two approaches are the best to reporting SPL experiments.
机译：为了获得价值，软件工程（SE）研究领域必须提供一定质量的结果。高质量的研究成果可以确保经验和知识，这对于将技术转让给行业至关重要。获得这种质量结果的一种方法是实验。实验是一种科学方法，旨在为建立因果关系的实际观察提供理论依据。进行良好，可审核且可重复的实验对于科学发展和新颖性至关重要。由于研究人员希望通过报告能够使实验重复的信息和读者能够理解实验并验证结果的方法来评估此类实验是否有所改进，因此文献中已讨论了SE中受控实验和准实验的质量评估。因此，这项工作从经验上比较了在软件产品线（SPL）上下文中SE实验质量评估的四种方法。此外，我们有兴趣通过讨论充分的重用技术（如SPL）来验证报告实验的质量。皮尔森技术支持评估方法对之间的相关性。此外，将T检验和Mann-Whitney-Wilcoxon U检验应用于样品，以验证使用实验模板时实验质量是否存在差异。初步结果显示它们之间有很强的正相关性，假设检验证实使用实验模板时质量存在这种差异，并且SPL实验报告的计划阶段比分析和解释阶段更多。根据我们的结果，我们提供了初步的证据，两种方法是报告SPL实验的最佳方法。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号