A new method of identifying influential nodes in complex software network based on leaderrank

作者：Huang Guoyan; Liu Jing^*; Chen Xiaojuan; Ren Jiadong

来源：ICIC Express Letters, 2016, 10(12): 2837-2844.

• Bi-directional links
• Dependency networks
• Influential nodes
• Leaderrank
• nocv1
• Software Evolution
• Software network
• Stability and robustness
• Weighted networks

Identifying influential nodes is crucial for understanding and improving the stability and robustness of complex software network. This paper presents a new method based on LeaderRank information to identify top-k influential nodes in directed-weighted complex software network. Firstly, the function and relationship of function calling or dependency are mapped to a Directed function dependency network (DN). Secondly, a so-called stem node connected with every other node by a bidirectional link is introduced into the original network. Thirdly, an algorithm Constructing DWN (C-DWN) is used to transform a DN into a Directed-Weighted function dependency network (DWN). Finally, the Identifying Influential Node in Software Network (IIN-SN) algorithm is proposed to identify influential nodes. Experimental results show that the proposed method is effective for identifying influential nodes with the final score ranking following power-laws. The top-k most influential nodes are the same and the final score (NS) of each node is a little larger during the software evolution. What is more, we present that some factors such as in-degree of nodes and influence of connected nodes are related to the influence of nodes.

• 出版日期2016