Application of self-organizing maps to software repositories in reuse-based software development

In reuse-based software development, developers create new software by using previously constructed software components. Successful software reuse requires the construction of a software repository to store, organize, and catalog collections of available software components. It is essential that the software repository not only be well structured but also provide efficient tools for developers to find reusable components that meet their needs. The self-organizing map (SOM), an unsupervised learning neural network, is a powerful data mining technique for clustering and visualization of huge data set. For clustering, SOM can distinguish input data into a number of clusters. For visualization, SOM can reduce a high-dimensional input space to a two-dimensional map. The resulting map can assist users to visualize both structures in the original data, and also the semantic relationships among them. In this paper, we study the concepts and methodology of SOM and its application to the organization of software repositories. We also examine several interesting SOM applications involving software repositories and report their distinctive features.