Extracting, identifiyng and visualisation of the content in software projects

Abstract

The paper proposes a method for extracting, identifying and visualisation of topics in software projects. In addition to standard information retrieval techniques, we use AST and WordNet ontology to enrich document vectors extracted from parsed source code, LSI to reduce its dimensionality and the swarm intelligence in the bee behaviour inspired algorithms to cluster documents contained in it. We extract topics from the identified clusters and visualise them in 3D graph. The goal is to provide insight into software projects for development participants in the process of analysing and reusing the source code.