eprintid: 17837
rev_number: 2
eprint_status: archive
userid: 1
dir: disk0/00/01/78/37
datestamp: 2023-12-19 03:24:09
lastmod: 2023-12-19 03:24:09
status_changed: 2023-12-19 03:08:46
type: article
metadata_visibility: show
creators_name: Muneer, A.
creators_name: Taib, S.M.
creators_name: Fati, S.M.
creators_name: Balogun, A.O.
creators_name: Aziz, I.A.
title: A hybrid deep learning-based unsupervised anomaly detection in high dimensional data
ispublished: pub
keywords: Clustering algorithms; Deep learning; Gradient methods; Multilayer neural networks; Optimization; Stochastic systems, Anomaly detection; Auto encoders; Critical researches; Deep learning; High dimensional data; Hybrid model; Optimization algorithms; Real-world problem; Research issues; Unsupervised anomaly detection, Anomaly detection
note: cited By 15
abstract: Anomaly detection in high dimensional data is a critical research issue with serious implication in the real-world problems. Many issues in this field still unsolved, so several modern anomaly detection methods struggle to maintain adequate accuracy due to the highly descriptive nature of big data. Such a phenomenon is referred to as the �curse of dimensionality� that affects traditional techniques in terms of both accuracy and performance. Thus, this research proposed a hybrid model based on Deep Autoencoder Neural Network (DANN) with five layers to reduce the difference between the input and output. The proposed model was applied to a real-world gas turbine (GT) dataset that contains 87620 columns and 56 rows. During the experiment, two issues have been investigated and solved to enhance the results. The first is the dataset class imbalance, which solved using SMOTE technique. The second issue is the poor performance, which can be solved using one of the optimization algorithms. Several optimization algorithms have been investigated and tested, including stochastic gradient descent (SGD), RMSprop, Adam and Adamax. However, Adamax optimization algorithm showed the best results when employed to train the DANN model. The experimental results show that our proposed model can detect the anomalies by efficiently reducing the high dimensionality of dataset with accuracy of 99.40, F1-score of 0.9649, Area Under the Curve (AUC) rate of 0.9649, and a minimal loss function during the hybrid model training. © 2022 Tech Science Press. All rights reserved.
date: 2022
publisher: Tech Science Press
official_url: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85117009228&doi=10.32604%2fcmc.2022.020732&partnerID=40&md5=0a89f42575a6dd3c06e57dbb05782837
id_number: 10.32604/cmc.2022.020732
full_text_status: none
publication: Computers, Materials and Continua
volume: 70
number: 3
pagerange: 6073-6088
refereed: TRUE
issn: 15462218
citation:   Muneer, A. and Taib, S.M. and Fati, S.M. and Balogun, A.O. and Aziz, I.A.  (2022) A hybrid deep learning-based unsupervised anomaly detection in high dimensional data.  Computers, Materials and Continua, 70 (3).  pp. 6073-6088.  ISSN 15462218