Research Article
Enhancement of Gravity Centrality Measure Based on Local Clustering Method by Identifying Influential Nodes in Social Networks
@INPROCEEDINGS{10.1007/978-3-031-18123-8_48, author={Pham Van Duong and Xuan Truong Dinh and Le Hoang Son and Pham Van Hai}, title={Enhancement of Gravity Centrality Measure Based on Local Clustering Method by Identifying Influential Nodes in Social Networks}, proceedings={Multimedia Technology and Enhanced Learning. 4th EAI International Conference, ICMTEL 2022, Virtual Event, April 15-16, 2022, Proceedings}, proceedings_a={ICMTEL}, year={2022}, month={10}, keywords={Influential node Node ranking Gravity model Local clustering K-shell SIR model Kendall’s Tau}, doi={10.1007/978-3-031-18123-8_48} }- Pham Van Duong
Xuan Truong Dinh
Le Hoang Son
Pham Van Hai
Year: 2022
Enhancement of Gravity Centrality Measure Based on Local Clustering Method by Identifying Influential Nodes in Social Networks
ICMTEL
Springer
DOI: 10.1007/978-3-031-18123-8_48
Abstract
Identifying influential nodes has great theoretical and practical implications in real-world scenarios such as search engines, social networks, and recommendation systems. Among the most essential issues in the field of complicated networks. Many approaches have been developed and deployed that have proven to be as effective as the gravity model. However, these models only focus on the local information of the node and ignore the information about the node neighbors or the global information of the network, leading to the gravity model is not really effective. This study focuses on improving the gravity model by considering the position information of the node based on the improvement of the k-shell decomposition algorithm. In addition, the article also uses the link of the node's neighbors by the local neighbor coefficient to increase the rigor for the local information of the node. The paper applies the SIR model to simulate the propagation effect of the node, then uses the Kendall Tau coefficient to evaluate the efficiency between the list of influence rankings. This research applies the monotonicity ratio to evaluate the resolution of the proposed ranking list. The efficiency of the recommended method is proven to outperform other methods on 5 social network datasets.
