An Improved Approach to Identifying Key Classes in Weighted Software Network

Abstract EN

To help the newcomers understand a software system better during its development, the key classes are in general given priority to be focused on as soon as possible.

There are numerous measures that have been proposed to identify key nodes in a network.

As a metric successfully applied to evaluate the productivity of a scholar, little is known about whether h -index is suitable to identify the key classes in weighted software network.

In this paper, we introduced four h -index variants to identify key classes on three open-source software projects (i.e., Tomcat, Ant, and JUNG) and validated the feasibility of proposed measures by comparing them with existing centrality measures.

The results show that the measures proposed not only are able to identify the key classes but also perform better than some commonly used centrality measures (the improvement is at least 0.215).

In addition, the finding suggests that mE-Weight defined by the weight of a node’s top k edges performs best as a whole.