Feature-Based Normality Models for Anomaly Detection

Detecting previously unseen anomalies in sensor data is a challenging problem for artificial intelligence when sensor-specific and deployment-specific characteristics of the time series need to be learned from a short calibration period. From the application point of view, this challenge becomes inc...

Full description

Saved in:
Bibliographic Details
Main Authors: Hui Yie Teh, Kevin I-Kai Wang, Andreas W. Kempa-Liehr
Format: Article
Language:English
Published: MDPI AG 2025-08-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/15/4757
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849405838675935232
author Hui Yie Teh
Kevin I-Kai Wang
Andreas W. Kempa-Liehr
author_facet Hui Yie Teh
Kevin I-Kai Wang
Andreas W. Kempa-Liehr
author_sort Hui Yie Teh
collection DOAJ
description Detecting previously unseen anomalies in sensor data is a challenging problem for artificial intelligence when sensor-specific and deployment-specific characteristics of the time series need to be learned from a short calibration period. From the application point of view, this challenge becomes increasingly important because many applications are gravitating towards utilising low-cost sensors for Internet of Things deployments. While these sensors offer cost-effectiveness and customisation, their data quality does not match that of their high-end counterparts. To improve sensor data quality while addressing the challenges of anomaly detection in Internet of Things applications, we present an anomaly detection framework that learns a normality model of sensor data. The framework models the typical behaviour of individual sensors, which is crucial for the reliable detection of sensor data anomalies, especially when dealing with sensors observing significantly different signal characteristics. Our framework learns sensor-specific normality models from a small set of anomaly-free training data while employing an unsupervised feature engineering approach to select statistically significant features. The selected features are subsequently used to train a Local Outlier Factor anomaly detection model, which adaptively determines the boundary separating normal data from anomalies. The proposed anomaly detection framework is evaluated on three real-world public environmental monitoring datasets with heterogeneous sensor readings. The sensor-specific normality models are learned from extremely short calibration periods (as short as the first 3 days or 10% of the total recorded data) and outperform four other state-of-the-art anomaly detection approaches with respect to F1-score (between 5.4% and 9.3% better) and Matthews correlation coefficient (between 4.0% and 7.6% better).
format Article
id doaj-art-ae095969651b46a1a80b6ab0e0b34b4e
institution Kabale University
issn 1424-8220
language English
publishDate 2025-08-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-ae095969651b46a1a80b6ab0e0b34b4e2025-08-20T03:36:34ZengMDPI AGSensors1424-82202025-08-012515475710.3390/s25154757Feature-Based Normality Models for Anomaly DetectionHui Yie Teh0Kevin I-Kai Wang1Andreas W. Kempa-Liehr2Department of Electrical, Computer and Software Engineering, The University of Auckland, Auckland 1142, New ZealandDepartment of Electrical, Computer and Software Engineering, The University of Auckland, Auckland 1142, New ZealandDepartment of Engineering Science and Biomedical Engineering, The University of Auckland, Auckland 1142, New ZealandDetecting previously unseen anomalies in sensor data is a challenging problem for artificial intelligence when sensor-specific and deployment-specific characteristics of the time series need to be learned from a short calibration period. From the application point of view, this challenge becomes increasingly important because many applications are gravitating towards utilising low-cost sensors for Internet of Things deployments. While these sensors offer cost-effectiveness and customisation, their data quality does not match that of their high-end counterparts. To improve sensor data quality while addressing the challenges of anomaly detection in Internet of Things applications, we present an anomaly detection framework that learns a normality model of sensor data. The framework models the typical behaviour of individual sensors, which is crucial for the reliable detection of sensor data anomalies, especially when dealing with sensors observing significantly different signal characteristics. Our framework learns sensor-specific normality models from a small set of anomaly-free training data while employing an unsupervised feature engineering approach to select statistically significant features. The selected features are subsequently used to train a Local Outlier Factor anomaly detection model, which adaptively determines the boundary separating normal data from anomalies. The proposed anomaly detection framework is evaluated on three real-world public environmental monitoring datasets with heterogeneous sensor readings. The sensor-specific normality models are learned from extremely short calibration periods (as short as the first 3 days or 10% of the total recorded data) and outperform four other state-of-the-art anomaly detection approaches with respect to F1-score (between 5.4% and 9.3% better) and Matthews correlation coefficient (between 4.0% and 7.6% better).https://www.mdpi.com/1424-8220/25/15/4757anomaly detectionfeature engineeringnormality modelsensor data qualitytime series analytics
spellingShingle Hui Yie Teh
Kevin I-Kai Wang
Andreas W. Kempa-Liehr
Feature-Based Normality Models for Anomaly Detection
Sensors
anomaly detection
feature engineering
normality model
sensor data quality
time series analytics
title Feature-Based Normality Models for Anomaly Detection
title_full Feature-Based Normality Models for Anomaly Detection
title_fullStr Feature-Based Normality Models for Anomaly Detection
title_full_unstemmed Feature-Based Normality Models for Anomaly Detection
title_short Feature-Based Normality Models for Anomaly Detection
title_sort feature based normality models for anomaly detection
topic anomaly detection
feature engineering
normality model
sensor data quality
time series analytics
url https://www.mdpi.com/1424-8220/25/15/4757
work_keys_str_mv AT huiyieteh featurebasednormalitymodelsforanomalydetection
AT kevinikaiwang featurebasednormalitymodelsforanomalydetection
AT andreaswkempaliehr featurebasednormalitymodelsforanomalydetection