Link prediction based on the powerful combination of endpoints and neighbors

Performance improvement of topological similarity-based link prediction models becomes an important research in complex networks. In the models based on node influence, researchers mainly consider the roles of endpoints or neighbors. Through investigations, we find that an endpoint with large influe...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:International journal of modern physics. B, Condensed matter physics, statistical physics, applied physics Condensed matter physics, statistical physics, applied physics, 2020-11, Vol.34 (28), p.2050269
Hauptverfasser: Gao, Tianrun, Zhu, Xuzhen
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Performance improvement of topological similarity-based link prediction models becomes an important research in complex networks. In the models based on node influence, researchers mainly consider the roles of endpoints or neighbors. Through investigations, we find that an endpoint with large influence has many neighbors. Meanwhile, the neighbors connect with more nodes besides endpoint, meaning that the endpoint can transmit extensive influence by the powerful combination of itself and neighbors. In addition, we evaluate the node influence by degree because the degree represents the number of neighbors accurately. In this paper, through focusing on the degree of endpoints and neighbors, we propose the powerful combination of endpoints and neighbors (PCEN) model. Experiments on twelve real network datasets demonstrate that the proposed model has better prediction performances than the traditional models.
ISSN:0217-9792
1793-6578
DOI:10.1142/S0217979220502690