Fine-grained Outlier Detection and Ranking

カテゴリ: 研究会(論文単位)

論文No: IS20046

グループ名: 【C】電子・情報・システム部門 情報システム研究会

発行日: 2020/10/12

タイトル(英語): Fine-grained Outlier Detection and Ranking

著者名: バックフース ヤナ(日立製作所),肥村 洋輔(日立製作所),増田 峰義(日立製作所),林 真一(日立製作所)

著者名(英語): Jana Backhus(Hitachi, Ltd. Research & Development Group),Yohsuke Himura(Hitachi, Ltd. Research & Development Group),Mineyoshi Masuda(Hitachi, Ltd. Research & Development Group),Shinichi Hayashi(Hitachi, Ltd. Research & Development Group)

キーワード: outlier detection|outlier ranking|network traffic|large-scale data|data context|root cause identification

要約(日本語): We propose an outlier detection and ranking framework to support large scale IT system operators’ manual monitoring efforts by splitting data into context-based data instances enabling fine-grained identification of system performance issues and its root

要約(英語): We propose an outlier detection and ranking framework to support large scale IT system operators’ manual monitoring efforts by splitting data into context-based data instances enabling fine-grained identification of system performance issues and its root causes. Evaluation shows identification of important root causes in proof-of concept real-world network traffic data.

原稿種別: 英語

PDFファイルサイズ: 1,030 Kバイト