Community Preserving Sign Prediction for Weak Ties of Complex Networks

Kangya He; Donghai Guan; Weiwei Yuan

Quality, Reliability, Security and Robustness in Heterogeneous Systems. 13th International Conference, QShine 2017, Dalian, China, December 16 -17, 2017, Proceedings

Research Article

Community Preserving Sign Prediction for Weak Ties of Complex Networks

Download

226 downloads

Cite: BibTeX Plain Text

@INPROCEEDINGS{10.1007/978-3-319-78078-8_2,
    author={Kangya He and Donghai Guan and Weiwei Yuan},
    title={Community Preserving Sign Prediction for Weak Ties of Complex Networks},
    proceedings={Quality, Reliability, Security and Robustness in Heterogeneous Systems. 13th International Conference, QShine 2017, Dalian, China, December 16 -17, 2017, Proceedings},
    proceedings_a={QSHINE},
    year={2018},
    month={4},
    keywords={Sign prediction Weak tie Link prediction Signed network},
    doi={10.1007/978-3-319-78078-8_2}
}

Kangya He
Donghai Guan
Weiwei Yuan
Year: 2018
Community Preserving Sign Prediction for Weak Ties of Complex Networks
QSHINE
Springer
DOI: 10.1007/978-3-319-78078-8_2

Kangya He¹^,*, Donghai Guan^,*, Weiwei Yuan^,*

1: Nanjing University of Aeronautics and Astronautics

*Contact email: kangyahe@gmail.com, dhguan@nuaa.edu.cn, yuanweiwei@nuaa.edu.cn

Abstract

The weak ties are crucial bridges between the tightly coupled node groups in complex networks. Despite of their importance, no existing work has focused on the sign prediction of weak ties. A community preserving sign prediction model is therefore proposed to predict the sign of the weak ties. Nodes are firstly divided into different communities. The weak ties are then detected via the connections of the divided communities. SVM classifier is finally trained and used to predict the sign of weak ties. Experiments held on the real world dataset verify the high prediction performances of our proposed method for weak ties of complex networks.

Keywords: Sign prediction Weak tie Link prediction Signed network

Published: 2018-04-25
Appears in: SpringerLink

: http://dx.doi.org/10.1007/978-3-319-78078-8_2