Estimating user’s intention and emotion by analyzing operation log data of IoT appliances

Abstract

Smartphones can be readily used to operate home appliances remotely, and we can gather log data when a user operates home appliances. However, currently, there is no established method for using the collected log data, and they remain unused. Therefore, assuming some relevance between the operation of home appliances and the intention and emotion of users, we aimed to establish a method for analyzing the former and understanding the latter. We implemented an application that can operate a cold/hot blower via smartphones while simultaneously surveying users’ intentions and emotions. It was used by participants daily for approximately 3 months. As a result, we found effective operation sequence rules for estimating intention and emotion and could construct a model that estimated intention and emotion with good accuracy from the operation log data using support vector machine.