Fig. 3

The workflow of Fog computing in the present study. Fog computing comprises three distinctive modes in a single time sequence: observation, machine learning, and status update. The observation phase executes streaming data storing. The Machine Learning phase does data processing and learning the model. The status update phase carries out a classification of a fan status with the trained model. A red box refers to a sliding window corresponding to a process of the workflow on the timeline. For example, the box which is on the second from the left indicates data storing. Likewise, the fifth box from the left is of learning the model