Fulltext available Open Access
License: 
Title: Deep learning for anomaly detection in multivariate time series data
Language: German
Authors: Assendorp, Jan Paul 
Keywords: Deep-Learning; Machine-Learning
Issue Date: 22-Nov-2017
Abstract: 
Das Erkennen von Anomalien in Sensordaten ist ein wichtiger Anwendungsfall in der Industrie, um Fehler in maschinellen Prozessen frühzeitig erkennen zu können und potentiellen Schäden vorzubeugen. In dieser Arbeit wird ein Deep-Learning-Verfahren entwickelt, welches in mehrdimensionalen Sensordaten ungewöhnliche Muster erkennen kann. Dafür werden Echtdaten aus einer industriellen Anwendung verwendet.

Anomaly detection is crucial for the procactive detection of fatal failures of machines in industry applications. This thesis implements a deep learning algorithm for the task of anomaly detection in multivariate sensor data. The dataset is taken from a real-world application.
URI: http://hdl.handle.net/20.500.12738/8162
Institute: Department Informatik 
Type: Thesis
Thesis type: Master Thesis
Advisor: von Luck, Kai 
Referee: Meisel, Andreas 
Appears in Collections:Theses

Files in This Item:
File Description SizeFormat
thesis_assendorp_2017.pdf3.03 MBAdobe PDFView/Open
Show full item record

Page view(s)

614
checked on Dec 25, 2024

Download(s)

1,560
checked on Dec 25, 2024

Google ScholarTM

Check

HAW Katalog

Check

Note about this record


Items in REPOSIT are protected by copyright, with all rights reserved, unless otherwise indicated.