基于测试成本元素的众包测试缺陷数量估计模型  

Defect Number Estimation Model for Crowdsourced Testing Based on Test Cost Elements

作　　者：姚奕 刘语婵 杨帆 YAO Yi;LIU Yu-chan;YANG Fan(Institute of Command and Control Engineering,Army Engineering University,Nanjing 210007,China)

机构地区：[1]陆军工程大学指挥控制工程学院,江苏南京210007

出　　处：《测控技术》2021年第9期7-15,31,共10页Measurement & Control Technology

基　　金：国家重点研发计划项目(2018YFB1403400);国家自然科学基金项目(61702544);中国博士后基金项目(2016M603031);江苏省自然科学基金项目(BK20160769,BK20141072)。

摘　　要：为实时监控众包测试任务过程从而对任务完成进行评估,针对分布式众包测试过程的不可预测性和测试成本的庞大性,从软件的可靠性出发,基于软件可靠性增长模型提出了一个同时考虑测试人力和测试报告两个测试成本元素的众包测试软件缺陷数量估计模型。首先分析两个成本元素的相关性,并依据成本元素相关性建立了一个分段式通用可靠性增长模型框架,并结合已有的3种测试工作量函数,以此估计出软件潜在缺陷数量和测试中的累积检测缺陷数量。在四组真实的众包测试数据集上的对比实验表明,模型的估计误差精度低于10%,优于传统可靠性增长模型。该模型能够利用较少实际数据进行缺陷预测,具有较强的实用性。In order to monitor the process of crowdsourced testing tasks in real time and evaluate the completion of tasks,in view of the unpredictability of the distributed crowdsourced testing process and the huge test cost,a software defect number estimation model for crowdsourced testing based on software reliability growth model,which considers the quantity of test labor and the number of test reports as two test cost elements,is proposed.Firstly,the correlation between two cost elements is analyzed,and a segmented general reliability growth model framework is established based on the correlation between the cost elements.Combined with the three existing test-effort functions,the number of potential software defects and the cumulative number of detected defects are estimated.Experiments in four groups of real crowdsourced test datasets show that the estimation error of defect number in the model is less than 10%,which is better than the traditional reliability growth model.This model can use fewer actual data for defect prediction and is proved to have strong practicability.

关 键 词：众包测试 测试成本元素 缺陷数量估计 软件可靠性增长模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]