Developer Modelling using Software Quality Metrics and Machine Learning

Software development has become an essential activity for organizations that increasingly rely on these to manage their business. However, poor software quality reduces customer satisfaction, while high-quality software can reduce repairs and rework by more than 50 percent. Software development is now seen as a collaborative and technology-dependent activity performed by a group of people. For all these reasons, choosing correctly software development members teams can be decisive. Considering this motivation, classifying participants in different profiles can be useful during project management team's formation and tasks distribution. This paper presents a developer modeling approach based on software quality metrics. Quality metrics are dynamically collected. Those metrics compose the developer model. A machine learning-based method is presented. Results show that it is possible to use quality metrics to model developers.