Mining and monitoring human activity patterns in smart environment-based healthcare systems

Abstract

Rapid evolution of sensing technology and increasing power computation has resulted in the emergence of smart environments with smart health services. Smart environments can generate hundreds of thousands of transactions per day and storage over the long term is a major issue. Therefore smart environment big data is utilized, which stores large volumes of datasets, both structured and unstructured. Healthcare services are recent and challenging aspects in analytics and sensor technology. People migrating from rural areas to urban areas affect healthcare services to a large extent. Due to migration and developing technology, cities around the world are investing in digital transformation, which aims to provide people with a healthier environment. New healthcare applications are based on activity recognition of people and can include wearable or ambient sensors to gather information related to health and human activity. Consumption of energy is analyzed along with the activity pattern of humans by determining the level of appliance usage that relates to human activity. The pattern determined is used to monitor elderly people living alone. The activity monitored can be learned just from appliances and their time usage. The proposed model detects appliance usage in particular with reliable accuracy. Mining is done purely based on absolute prediction along with clustering models.