Software typically consists of a large number of components (in software design terms called ‘classes’). Master student Xavyr Rademaker worked on new ways of automatically determining the role of such a software component using a combination of machine learning and complex network analysis. This can be usefil, as it may allow poorly documented software projects to be understood more efficiently by software developers working on the project. Xavyr will present his findings at the Complex Networks conference in Cambridge in December 2018. The work is summarized in the following paper:

• X. Rademaker, M.R.V. Chaudron and F.W. Takes, Automatic identification of component roles in software design networks, in Proceedings of the 7th International Conference on Complex Networks, Studies in Computational Intelligence 813: 145-157, Springer, 2018.