Classification of Sleep Patterns by Means of Markov Modeling and Correspondence Analysis
Ben H. Jansen, Weikang Cheng · IEEE Transactions on Pattern Analysis and Machine Intelligence · 1987
Shown is how correspondence analysis can be used to track changes in an individuals' sleep pattern. Correspondence analysis was applied to sleep stage transition matrices computed from all-night sleep of normal, obese, and apnoetic subjects. Differences between the groups, and intraindividual changes in sleep patterns could be visualized better than with a x2-based clustering approach.