Discovering Common Features in Software Code Using Self-Organizing Maps

The self-organizing map is discussed as an unsupervised clustering method. Its ability to form clusters indicates similar features in a data set. Based on this property, it is demonstrated that a self-organizing map is capable of identifying features within software code by grouping procedures with similar properties together. This allows us to identify potential objects, abstract data types or classes. In experiments with a simulation package Pascal SIM (a procedural oriented implementation) as the data set, features were identified and a feature matrix constructed that served as the input to the self-organizing map. The results obtained were clusters on the map that indicated procedures with similar features being grouped together. This demonstrates that the self-organizing map is potentially a viable tool in intelligently automating the discovery of common features and groupings within code.