Enhancing anomaly detection using temporal pattern discovery

Technological enhancements aid development and research in smart homes and intelligent environments. The temporal nature of data collected in a smart environment provides us with a better understanding of patterns that occur over time. Predicting events and detecting anomalies in such data sets is a complex and challenging task. To solve this problem, we suggest a solution using temporal relations. Our temporal pattern discovery algorithm, based on Allen's temporal relations, has helped discover interesting patterns and relations from smart home data sets. We hypothesize that machine learning algorithms can be designed to automatically learn models of resident behavior in a smart home and, when these are incorporated with temporal information, the results can be used to detect anomalies. We describe a method of discovering temporal relations in smart home data sets and applying them to perform anomaly detection on the frequently occurring events by incorporating temporal relation information shared by the activity. We validate our hypothesis using empirical studies based on the data collected from real resident and virtual resident (synthetic) data.