Detection and Analysis of behavior change based on unsupervised clustering method for smart home technology

Smart home environments offer a novel opportunity to indubitably monitor human behavior. Data collected from smart home sensors can form a time series in which each data point is paired with an associated timestamp. For human health, this time-series data is valuable when detecting and analyzing changes associated with seasonal variations, new lifestyle choices, new symptoms of the disease, and so on. To detect and analyze behavior changes that accompany health events, we study unsupervised clustering approach. We use two unsupervised algorithms, Sliding Window and SWAB clustering, to detects activity timing and duration changes between windows of time, determines the significance of the detected changes, and analyzes the nature of the changes. We demonstrate our approach using a case study for an older adult living in smarthome who experienced major health events, including insomnia. Our algorithms detect behavior changes consistent with the medical literature for this case. The results suggest the changes can be automatically detected based on significance changes using these the two algorithms. The proposed change detection approach is useful to understanding the behavioral effects of major health conditions

Keywords:

Smart Home , Smart Health , Behavior Change Detection , Clustering