A cognitive model to predict human interest in smart environments

Abstract

Recently, the idea of smart cities has made several strides forward in literature. Work has hypothesize that the combination of Artificial Intelligence, Cloud Computing, and High powered computers will make technology more human-centric, even, the idea that smart cities will be able to understand the thought process of a human being seems very much likely today. This paper is along this line of thought. In particular, we try to present a method to model the cognitive state of human interest. This is done to take one more step towards the realization of a smart cognitive city. An approach which is Subjective–Objective in nature is presented to model the computation of activity inspired by interest. Based on activity, human latent state values are indirectly deduced. Inspiration is drawn from Physics and interest is modeled upon the Ornstein–Uhlenbeck (OU) process. Concepts of Adaptive filtering are used to formulate an evolving transformation function that automatically and adaptively models the conversion of interest into activity. Particle filter is employed to provide an elucidation which is computationally feasible. To validate the viability of the method, experimentation is performed with real datasets. © 2020 Elsevier B.V.