Assistive Dementia Care System Through Smart Home

As per the statistics of World Health Organization, a major percentage of elderly society across the globe is affected with dementia (age related memory loss). Dementia care entails prolonged effort in terms of money, time and manpower. Assistive health care system is therefore essential and is probable through a smart home, that offers Ambient Assisted Living (AAL) to its occupants. The objective of this work is to model an Intelligent Decision Support System (IDSS) for dementia care through the smart home. The innovation in the design of IDSS is to offer two levels of decision-making (1) Short-Term Decision-Making (STDM)—to raise suitable alerts for the abnormality detected in ADL (2) Long-Term Decision-Making (LTDM)—to decide on the progress in occupant’s developmental stage of dementia. The novelty in the design of STDM is to assimilate Random Forest (data driven) decision-making and Rule-based (knowledge-driven) decision-making within a single framework. Random Forest modeling provides better predictive accuracy through ensemble learning which is later combined with the domain specific knowledge to offer context-based decision-making. On the other hand, LTDM decides on occupants developmental stage of dementia through automation of Barthel score. Barthel score is a clinical measure to assess the stage of dementia through the level of dependency required by the occupant to complete his activities. The experimental analysis confirms the proficiency of the proposed IDSS in decision-making is better than existing approaches.