Inference of activities with unexpected actions using pattern mining
Shamila Nasreen, Muhammad Awais Azam, Usman Naeem, Mustansar Ali Ghazanfar · 2015
Recognition of activities in an unobtrusive manner has attracted the attention of context aware systems, which provide end users with services based on everyday activities that are recognised without infringing the privacy of the end user. Current work has generally focused on applying a range of traditional classification and semantic reasoning based techniques in order to recognise these activities. However, the ability to recognise unexpected actions while the activity is being conducted remains a challenge. In this paper, we present an approach that is able to recognise activities regardless of the order of tasks/actions used to perform the activity. The proposed recognition framework extends an existing activity recognition approach by deploying a frequent pattern mining technique to find patterns among different streams of captured sensor events in order to increase the adaptive learning of the proposed recognition approach.