Just-in-time Adaptive Anomaly Detection and Personalized Health Feedback

The rapid population aging and the availability of sensors and intelligent objects motivate the development of information technology-based healthcare systems that meet the needs of older adults by supporting them to continue their day-to-day activities. These systems collect information regarding the daily activities of the users that potentially helps to detect any significant changes and to provide them with relevant and tailored health-related information and quality of life-improving suggestions. To this aim, we propose a Just-in-time adaptive intervention system that models the user daily routine using a task model specification and detects relevant contextual events that occurred in their life in order to detect anomalous behaviors and strategically generate tailored interventions to encourage behaviors conducive to a healthier lifestyle. The system uses a novel algorithm to detect anomalies in the user daily routine. In addition, by a systematic validation through a system that automatically generates wrong sequences of activities, we show that our anomaly detection algorithm is able to find behavioral deviations from the expected behavior at different times along with the category of the anomalous activity performed by the user with good accuracy. Later, the system uses a Mamdani-type fuzzy rule-based component to predict the level of intervention needed for each detected anomaly and a sequential decision-making algorithm, Contextual Multi-armed Bandit, to generate suggestions to minimize anomalous behavior. We describe the system architecture in detail, and we provide example implementations for corresponding health feedback. To test our approach, we collected sensor data in our smart lab testbed while an actor was performing activities of daily living over a period of 2 weeks.