A similarity measure approach for identifying causes of anomaly in activities of daily living

Anomaly detection in Activities of Daily Living is a challenging task driven by the need to improve the quality of life and promote independent living of the increasing ageing population. There are many computational methodologies for detecting anomalies. They are mainly based on the concept of learning usual activities of daily living routines and detect abnormalities in it. However, they are limited by their inability to predict the actual cause of the anomaly. Understanding the cause of the anomalies can enable robust anomaly detection system to be built with a low rate of false alarms. This paper proposes a similarity measure approach for identifying the cause of anomalies in activities of daily living routine. The proposed approach is based on a pair-wise similarity measure of the features present in a dataset. Preliminary experiments conducted on both real and synthetic data achieve an excellent result with an overall accuracy of 96%.