An unsupervised learning method for perceived stress level recognition based on office working behavior

Abstract

The health issues in office workers regarding of working environment and working behavior have raised many concerns, both in medical field and technological field. For medical field, the concerns were related to physical injuries and stress due to either bad environment or bad working behaviors. In technological field, the main concern was to find a proper solution to prevent and raise awareness to these issues. In this paper, we discussed the possibility of using unsupervised learning for clustering office working behavior to show the relationship of the working behavior and stress level. We used the data collected from the device which include both behavior data and environment data. The results successfully demonstrated the two clusters that represents the working behavior related to either high or low stress level. The results can be used further to develop a classification model and to raise awareness in office workers.