Presentation | 2022-05-12 Why Do Small Eigenvalues of Laplacian Matrix Improve Anomaly Detection of Temporal Networks? Eriko Segawa, Yusuke Sakumoto, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Many real networks are temporal networks in which the nodes and their relationships change over time, and technology to detect anomalies is needed for such networks. LAD~(Laplacian Anomaly Detection) is an innovative method to detect anomalies in temporal networks using the large eigenvalues of the Laplacian matrix that represents the network structure. In previous work, we clarified that the success rate of anomaly detection by LAD improved significantly using a combination of small and large eigenvalues, but the reason for this has not been fully elucidated. In this paper, we discuss why the combination of large and small eigenvalues is useful for anomaly detection through experiments using temporal networks with various types of anomalies. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Anormaly Detection / Dynamic Network / Spectral Graph Theory / Laplacian Matrix / Social Network Analysis |
Paper # | CQ2022-9 |
Date of Issue | 2022-05-05 (CQ) |
Conference Information | |
Committee | CQ / CS |
---|---|
Conference Date | 2022/5/12(2days) |
Place (in Japanese) | (See Japanese page) |
Place (in English) | Fukui (Fuku Pref.) |
Topics (in Japanese) | (See Japanese page) |
Topics (in English) | Optical/Wireless Access and Their Integration, Communication Behavior, QoE and Psychology, Assessment / Measurement / Control / Optimization of Communication Quality, Network Services, Wireless Networks, MIMO/Diversity/Multiplexing Techniques, etc. |
Chair | Jun Okamoto(NTT) / Jun Terada(NTT) |
Vice Chair | Takefumi Hiraguri(Nippon Inst. of Tech.) / Gou Hasegawa(Tohoku Univ.) / Daisuke Umehara(Kyoto Inst. of Tech.) |
Secretary | Takefumi Hiraguri(NTT) / Gou Hasegawa(Ritsumeikan Univ.) / Daisuke Umehara(NICT) |
Assistant | Yoshiaki Nishikawa(NEC) / Ryoichi Kataoka(KDDI Research) / Kimiko Kawashima(NTT) / Takahiro Yamaura(Toshiba) / Yuta Ida(Yamaguchi Univ.) |
Paper Information | |
Registration To | Technical Committee on Communication Quality / Technical Committee on Communication Systems |
---|---|
Language | JPN |
Title (in Japanese) | (See Japanese page) |
Sub Title (in Japanese) | (See Japanese page) |
Title (in English) | Why Do Small Eigenvalues of Laplacian Matrix Improve Anomaly Detection of Temporal Networks? |
Sub Title (in English) | |
Keyword(1) | Anormaly Detection |
Keyword(2) | Dynamic Network |
Keyword(3) | Spectral Graph Theory |
Keyword(4) | Laplacian Matrix |
Keyword(5) | Social Network Analysis |
1st Author's Name | Eriko Segawa |
1st Author's Affiliation | Kwansei Gakuin University(Kwansei Gakuin Univ.) |
2nd Author's Name | Yusuke Sakumoto |
2nd Author's Affiliation | Kwansei Gakuin University(Kwansei Gakuin Univ.) |
Date | 2022-05-12 |
Paper # | CQ2022-9 |
Volume (vol) | vol.122 |
Number (no) | CQ-15 |
Page | pp.pp.44-49(CQ), |
#Pages | 6 |
Date of Issue | 2022-05-05 (CQ) |