مريم جواهري منش

Many real-world systems can be modeled as complex networks to facilitate the study of systems by analysis of the complex network. In a complex network, some nodes can affect the performance of the entire network more than other nodes. These nodes are called important nodes and their number of them is low. In complex network studies, ranking and identifying important nodes is one of the most fundamental issues. Identifying important nodes has attracted the attention of many researchers in recent years. Global feature-based methods are unsuitable for large-scale complex networks due to their high time complexity. Methods based on local features, such as degree only examine the information of close neighbors. In addition, compared to methods that consider only one feature, considering multiple features can increase the performance of the method used. So, this work proposes a weighted centrality with multiple local features based on information entropy, degree composition, and clustering coefficient. Neighborhood information gets used to eva‎luate the influence of nodes and identify important nodes in complex networks. Degree and clustering coefficient are chosen as criteria. Then, information entropy is used to weigh the important measure, and the value of each node is obtained by calculating the weighted sum of these measures. Finally, all the nodes are ranked based on the obtained value and the important nodes can be identified. The proposed method for identifying important nodes gets applied to ten real networks and three artificial networks and gets compared with three basic methods and two new methods. The experimental results of the performance of the proposed method in rating differentiation show that in the networks with a specific structure, compared to the similar works is improved, and it has less time complexity comparing to other similar work.

شبكه‌هاي پيچيده , گره‌هاي مهم , معيار مركزيت , گراف‌كاوي

Complex Networks , Important nodes , centrality measure , Graph mining

Maryam Javaherimanesh

Dr. Mohammad Abdollahi