Elderly home care anomaly detection

Abstract

In this article, we provide a methodology for identifying irregularities in senior home care. In order to spot abnormalities, two situations are looked into: 1) the old person's vital signs and their environment; and 2) their movement patterns. Utilizing the isolation forest, which uses an isolation method to find anomalies on a random forest of decision trees, we assessed our suggested model. We contrast statistical approaches on labelled data with isolation forest on unlabeled data. Then, we contrast the isolation notion with a distance measure concept in order to demonstrate its dependability. The experiment demonstrates that isolation forest has higher detection accuracy and lower error prediction for two attributes in the first scenario: skin temperature and heart rate, while multi-covariance determinant has higher detection accuracy and lower error prediction in the second scenario.