Presentation 2022-11-05
Smoothing methods for reducing false positives in performance anomaly detection using machine learning
Taku Wakui, Mineyoshi Masuda,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In the integrated infrastructure management market, ML-based anomaly detection is one of the key features. However, computational speed of our existing anomaly detection is a little slow to apply to large VM environments. Therefore, it is necessary to significantly improve the computational speed. Although the computational speed can be improved by using a smaller ML model, reducing the model size adversely affects the accuracy of anomaly detection. In this report, we describe about two original smoothing methods which are applied to the output of the ML model and improve the anomaly detection accuracy of ML model whose model size has been reduced for speed.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) IT Operations / AIOps / Anomaly Detection / Machine Learning
Paper # KBSE2022-41,SC2022-36
Date of Issue 2022-10-28 (KBSE, SC)

Conference Information
Committee KBSE / SC
Conference Date 2022/11/4(2days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Takuya Saruwatari(NTT Data) / Kosaku Kimura(Fujitsu Lab.)
Vice Chair Yoshinori Tanabe(Tsurumi Univ.) / Takao Nakaguchi(KCGI)
Secretary Yoshinori Tanabe(Osaka Inst. of Tech.) / Takao Nakaguchi(NAIST)
Assistant Yoshitaka Aoki(BIPROGY) / Hiroki Horita(Ibaraki Univ.) / Shigeru Hosono(Tokyo Univ. of Tech.)

Paper Information
Registration To Technical Committee on Knowledge-Based Software Engineering / Technical Committee on Service Computing
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Smoothing methods for reducing false positives in performance anomaly detection using machine learning
Sub Title (in English)
Keyword(1) IT Operations
Keyword(2) AIOps
Keyword(3) Anomaly Detection
Keyword(4) Machine Learning
1st Author's Name Taku Wakui
1st Author's Affiliation Hitachi, Ltd.(Hitachi)
2nd Author's Name Mineyoshi Masuda
2nd Author's Affiliation Hitachi, Ltd.(Hitachi)
Date 2022-11-05
Paper # KBSE2022-41,SC2022-36
Volume (vol) vol.122
Number (no) KBSE-238,SC-239
Page pp.pp.60-65(KBSE), pp.60-65(SC),
#Pages 6
Date of Issue 2022-10-28 (KBSE, SC)