Skip to Main Content
One of key components in the development of smart home technology is the detection and recognition of activities of daily life. Based on a self-adaptive neural network called growing self-organizing maps (GSOM), this paper presents a new computational approach to cluster analysis of human activities of daily living within smart home environment. It was tested on a dataset collected from a set of simple state-change sensors installed on a one-bedroom apartment during a period of about two weeks. The results obtained indicate that, due to its advanced evolving, self- adaptive properties, the GSOM exhibits several appealing features in the analysis of useful patterns encoded in daily activity data. The approaches described in this paper contribute to the development of a user-friendly and interactive data-mining platform for the analysis of human activities within smart home environment through the improvement of pattern discovery, visualization and interpretation.