AI4SE：通过工具支持的分类获得新见解——两个工业案例.pdf

1、Leveraging AI for the classification of SE artifacts-industrial cases working with defect reportsWork by:Muhammad Laiq,Nauman bin Ali,Emelie Engstrm and Jrgen Brstlernauman.alibth.seNAUMAN BIN ALIo Senior lecturer(Docent)at Blekinge Institute of Technology,(BTH)Karlskronao Moved to Vallentuna(30 kms

2、 north)o Industry collaborators:Research interests-Test case and suite quality-Test case selection and prioritization-Test flakiness-Visual and data analytics for decision support-Levering test results for decisions about:-test scoping-release readiness-Anomaly detection in cloud based products and

3、servicesDefect reportDescription of a mismatch between the experienced and expected behavior of a systembug report,trouble report,issue report,failure report,error report,incident report,problem logNot all defect reports are equalType of solution required:Code correction to fix an error in the imple

4、mentationDocumentation update to fix an error/clarification in the description of featuresTraining and education to improve awareness of the systemsNew feature development requires a decision where the functionality should be added to the systemWho should be involved in the investigation and resolut

5、ion of the bug report,and when?Herzig,Kim,Sascha Just,and Andreas Zeller.Its not a bug,its a feature:how misclassification impacts bug prediction.2013 35th international conference on software engineering(ICSE).IEEE,2013.Defect report classificationWhy categorize:TaskClassificationShould developers

6、be involved?Defect classification based on the type of resolution required.A solution may require a code change,customer or support personnel education,an update to the documentation,or requesting a new feature.Not straight forwardThere are several review steps to filter and sort the defect reports