Similarity Recognition of Interval-Based Sleep Data

Over the last years the number of patients with sleep or sleep-related disorders is continuously growing. Not only sleep laboratories are able to monitor the sleep of patients but also consumer devices like smartphones or fitness trackers allow sleep recording to everyone. A drawback of professional as well as consumer recording is the hardly researched field of comparing similarities within sleep data. This paper presents a novel approach that allows the recognition of similarities between different sleep data sets based on extracted time intervals like sleep stages. The evaluation of the first proof of concept shows its suitability to distinguish between similar and dissimilar sleep data sets. The results open the door for further optimizations of the underlying approach and for further studies e.g. anomaly detection in medical data.