University of Lancaster and Howz release further change detection methodologies

There is clear evidence of the link between behaviour and wellbeing. Behavioural disturbances, such as agitation, wandering and sleep disorder, and changes in routine and environment, are associated with a higher risk of hospitalisation for vulnerable adults.

Analysing the daily routine and observing the night-time sleep pattern play an important role in monitoring health and wellbeing.

By using IoT sensors within a home, residents' behaviours can be understood and anomalies were detected. This allows care givers to provide support where it is needed.

Within this process, detecting change is important since it allows us to detect potential health problems. Our ability to detect breaks in regular routines allow us to provide the best possible insights to health and social care.

But how does Howz detect understand what should and shouldn’t be considered a change?

Howz has worked closely with Lancaster University to build a scientific approach to detecting change at scale. Together we have collaborated on multiple studies and this latest paper shares the new methodologies for change detection that we have now released into Howz.

The approach described in this paper is further step for our technology, adding to algorithms already deployed.

The problem of health and care of people is being revolutionised. An important component of that revolution is disease prevention and health improvement from home. A natural approach to the health problem is monitoring changes in people’s behavior or activities. These changes can be indicators of potential health problems. However, due to a person’s daily pattern, changes will be observed throughout each day, with, eg, an increase of events around meal times and fewer events during the night.

We do not wish to detect such within-day changes but rather changes in the daily behavior pattern from one day to the next. To this end, we assume the set of event times within a given day as a single observation. We model this observation as the realization of an inhomogeneous Poisson process where the rate function can vary with the time of day. Then, we propose to detect changes in the sequence of inhomogeneous Poisson processes. This approach is appropriate for many phenomena, particularly for home activity data. Our methodology is evaluated on simulated data. Overall, our approach uses local change information to detect changes across days. At the same time, it allows us to visualize and interpret the results, changes, and trends over time, allowing the detection of potential health decline.

Read the paper here