Goal-Oriented Learning in Experimental Software Engineering by Using Rough Sets

Rough set theory is a foundation for a series of analysis methods which has been successfully applied in many real-life problems of various areas. Main focus of the approach is to explain experimental data by establishing cause-effect relationships. We have applied the approach in the context of experimental software engineering to support analysis and interpretation of measurement data within interactive feedback sessions. Measurement was performed following the Goal/Question/Metric-approach. As a case study, we report experiences from a project at Allianz Lebensversicherungs-AG devoted to better understanding of and subsequent improvement of flexibility in software development. Experiences are reported from both application and methodological point of view.