Methodology for forecasting user experience for smart and assisted living in affect aware systems

Abstract

The application of affect aware systems in a smart home environment has the potential to address challenges in creation of a smart, adaptive, context-aware and assistive living space for elderly people. Affect aware systems that can predict the user experience even before an activity is performed, would provide a scope for effective communications and interactions between users and systems to ensure an adaptive and ambient living environment thereby ensuring more trust on intelligent systems. This paper therefore proposes an approach for forecasting user experiences in the context of affect aware systems to assist elderly people in smart homes. The preliminary result of the proposed analytical approach is supported by a random-forest based predictive model that achieves an overall prediction accuracy of 67% when evaluated on a subset of the UK DALE Dataset. [1]