Evaluating a patient’s ability to perform daily activities is critical to successful nursing and healthcare, particularly in the elderly. Such an assessment is a powerful predictor of so-called morbidity, or how much the patient is affected detrimentally by their symptoms. New research in the International Journal of Ad Hoc and Ubiquitous Computing is looking to develop a machine learning approach that can address the task of recognising a patient’s different activities in a smart home.
Salima Sabri and Abdelouhab Aloui of the Université de Bejaia in Algeria have evaluated their approach by comparing it to a Markov statistical approach and using several performance measures over three datasets. “We show how our model achieves significantly better recognition performance on certain data sets and with different representations and discretisation methods with an accuracy measurement that exceeds 92% and accuracy of 68%,” they report.
The team explains how context-aware systems are coming to the fore in healthcare research for monitoring the negative symptoms of an aging population without the need for undue invasiveness on the part of healthcare workers. Classification based on well-known and well-established indicators might now be incorporated into an automated system to show how well a patient can care for themselves or whether intervention is needed to assist them in coping with their symptoms.
Sabri, S. and Aloui, A. (2019) ‘A new approach for the recognition of human activities’, Int. J. Ad Hoc and Ubiquitous Computing, Vol. 32, No. 4, pp.211–223.