arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.04651cs.SE
收藏

引入集成机器学习算法以实现基于学习的测试中的自动测试用例生成

Introducing Ensemble Machine Learning Algorithms for Automatic Test Case Generation using Learning Based Testing

Sheikh Md. Mushfiqur Rahman, Nasir U. Eisty

AI总结:

本文研究在基于学习的测试中结合集成方法与基础分类器进行模型推断,实验证明Boosting生成故障检测测试用例的总体效果优于随机生成。

摘要:

集成方法是功能强大的机器学习算法,通过组合多个模型来增强预测能力并降低泛化误差。然而,它们在为被测系统(SUT)生成可有效检测故障的测试用例方面的潜力尚未得到广泛研究。本研究旨在系统考察在基于学习的测试(LBT)算法中,将集成方法与基础分类器相结合用于模型推断,从而为SUT生成可检测故障的测试用例,作为概念验证。我们针对若干函数开展了一系列实验,在所提出的LBT方法中使用不同的集成方法与分类器组合进行模型推断,以生成有效测试用例。随后,我们依据变异评分对各测试套件进行比较。结果表明,Boosting集成方法在生成有效测试用例方面总体表现更佳,并且所提方法的表现优于随机生成。该分析有助于为不同类型的函数确定合适的集成方法。通过将集成方法纳入LBT,本研究有助于理解如何利用集成方法实现有效的测试用例生成。

英文摘要:

Ensemble methods are powerful machine learning algorithms that combine multiple models to enhance prediction capabilities and reduce generalization errors. However, their potential to generate effective test cases for fault detection in a System Under Test (SUT) has not been extensively explored. This study aims to systematically investigate the combination of ensemble methods and base classifiers for model inference in a Learning Based Testing (LBT) algorithm to generate fault-detecting test cases for SUTs as a proof of concept. We conduct a series of experiments on functions, generating effective test cases using different ensemble methods and classifier combinations for model inference in our proposed LBT method. We then compare the test suites based on their mutation score. The results indicate that Boosting ensemble methods show overall better performance in generating effective test cases, and the proposed method is performing better than random generation. This analysis helps determine the appropriate ensemble methods for various types of functions. By incorporating ensemble methods into the LBT, this research contributes to the understanding of how to leverage ensemble methods for effective test case generation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑