抄録
In this paper, we introduce a real nursing sensor dataset which includes labeled dataset for supervised machine learning and the big data combined with patient medical records and sensors attempted for 2 years, and also describe a method for recognizing activities for a whole day utilizing prior knowledge about the activity segments in a day and utilizing importance sampling and Bayesian estimation, based on our paper at UbiComp2015 [13]. Moreover, we demonstrate data mining by applying our method to the bigger data with additional hospital data. Our method of recognizing a whole day activities outperformed the traditional method without using the prior knowledge. Moreover, the method significantly reduces the duration errors of activity segments. We also demonstrate a data mining applying our method to bigger data in a hospital, and show several results about the correlations with nurse profiles and patients status using Random Forest regression.
