Software modularization using Combination of Multiple Clustering

Clustering is a useful technique to group similar data entities based on their features. Clustering uses similarity or distance measures to find groups among entities. Many clustering algorithms and measures have been proposed in the literature. These algorithms and measures have their own strengths and weaknesses in the software clustering domain. To combine the strengths of various algorithms/measures, researchers have inte- grated more than one algorithm/measure in a single clustering process. This approach allowing cooperation between actors, is called Combination of Multiple Clustering approach (CMC). Although the use of CMC has been explored in various disciplines, little work has been done using CMC in the software domain. In this paper, we explore the idea of Cooperative Clustering (CC), a type of CMC, for software modularization. Our CC combines the strengths of similarity and distance measures together in a single clustering process. Software modularization is very important for architectural understanding. Modularization is the breaking down of a software system into modules so that similar entities (e.g. classes or functions) are collected together. We expect high quality results of CC in terms of authoritativeness. which is very important for architectural understanding.