SYSTEM AND METHOD FOR EQUIPMENT ABNORMALITY DIAGNOSIS BASED ON MULTIPLE TRANSFER ACTIVE LEARNING

The present invention relates to a system for diagnosing an abnormality of equipment based on multi-transition active learning and a method thereof. The system for diagnosing an abnormality of equipment based on multi-transition active learning according to the present invention includes: an abnorma...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: SON SEUNG CHUL, LEE BYUNG TAK, KO SEOK KAP, KIM EDEN
Format: Patent
Sprache:eng ; kor
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The present invention relates to a system for diagnosing an abnormality of equipment based on multi-transition active learning and a method thereof. The system for diagnosing an abnormality of equipment based on multi-transition active learning according to the present invention includes: an abnormality diagnosis module which checks whether there is an abnormality by examining collected data; a multi-transition module which generates a plurality of active learning models by using the abnormality diagnosis module; and a labeling module which performs a labeling task by presenting label request target data extracted by using an active learning model, and additionally trains the abnormality diagnosis module by using new label data. 본 발명은 다중전이 능동학습 기반 설비 이상 진단 시스템 및 그 방법에 관한 것이다. 본 발명에 따른 다중전이 능동학습 기반 설비 이상 진단 시스템은 수집 데이터를 검토하여 이상 여부를 확인하는 이상 진단 모듈과, 이상 진단 모듈을 이용하여 복수 개의 능동 학습 모델을 생성하는 다중 전이 모듈 및 능동 학습 모델을 이용하여 추출된 레이블 요청 대상 데이터를 제시하여 레이블링 작업을 수행하고, 새로운 레이블 데이터를 이용하여 이상 진단 모듈을 추가 학습시키는 레이블링 모듈을 포함한다.