Research Article
Identification of Abnormal Behavior in Activities of Daily Life Using Novelty Detection
@INPROCEEDINGS{10.1007/978-3-031-34776-4_29, author={Mauricio Freitas and Vinicius de Aquino Piai and Rudimar Dazzi and Raimundo Teive and Wemerson Parreira and Anita Fernandes and Ivan Miguel Pires and Valderi Reis Quietinho Leithardt}, title={Identification of Abnormal Behavior in Activities of Daily Life Using Novelty Detection}, proceedings={Mobile and Ubiquitous Systems: Computing, Networking and Services. 19th EAI International Conference, MobiQuitous 2022, Pittsburgh, PA, USA, November 14-17, 2022, Proceedings}, proceedings_a={MOBIQUITOUS}, year={2023}, month={6}, keywords={Novelty Detection Anomaly Detection Activities of Daily Living Machine Learning One-Class Support Vector Machine (OC-SVM) Local Outlier Factor (LOF)}, doi={10.1007/978-3-031-34776-4_29} }- Mauricio Freitas
Vinicius de Aquino Piai
Rudimar Dazzi
Raimundo Teive
Wemerson Parreira
Anita Fernandes
Ivan Miguel Pires
Valderi Reis Quietinho Leithardt
Year: 2023
Identification of Abnormal Behavior in Activities of Daily Life Using Novelty Detection
MOBIQUITOUS
Springer
DOI: 10.1007/978-3-031-34776-4_29
Abstract
The world population is aging at a rapid pace. According to the WHO (World Health Organization), from 2015 to 2050, the proportion of elderly people will practically double, from 12 to 22%, representing 2.1 billion people. From the individual’s point of view, aging brings a series of challenges, mainly related to health conditions. Although, seniors can experience opposing health profiles. With advancing age, cognitive functions tend to degrade, and conditions that affect the physical and mental health of the elderly are disabilities or deficiencies that affect Activities of Daily Living (ADL). The difficulty of carrying out these activities within the domestic context prevents the individual from living independently in their home. Abnormal behaviors in these activities may represent a decline in health status and the need for intervention by family members or caregivers. This work proposes the identification of anomalies in the ADL of the elderly in the domestic context through Machine Learning algorithms using the Novelty Detection method. The focus is on using available ADL data to create a baseline of behavior and using new data to classify them as normal or abnormal daily. The results obtained using the E-Health Monitoring database, using different Novelty Detection algorithms, have an accuracy of 91% and an F1-Score of 90%.
