In this paper, we present Hidden Markov Models (HMM) approach for forecasting the changes of sleep. Sleep is a major part of our life, and the amount and quality of sleep are closely related to our health. Forecasting changes of sleep is equivalent to forecasting changes of the health. We use numerous HMM models that is trained by datasets clustered on similarity basis. We find the optimal models with best probabilities in various learned HMM models and use this model to predict next sleep state. The sleep data are collected by Fitbit-HR from 150 healthy persons. The prediction performance was accuracy = 68.53% and recall = 68.19%
Research Interests
Digital Phenotyping; Data Mining & Machine Learning for Text & Multimedia; Brain-Sense-ICTConvergence Computing; Computational Olfaction Measurement; Simulation&Modeling