| Keywords |
Occupant status detection ; Machine learning ; Classification algorithm |
| Abstract |
The purpose of this study is to develope an occupant status detection model by using indoor environmental data such as temperature, humidity, CO2, noise, lighting power energy usage, etc. This study tested various classification algorithms (i.e., Support Vector Machine(SVM), K-Nearest Neighbor(KNN), Decision Trees(DT)) which are one of machine learning methods. We defined the occupant's state as 'Away', 'Active', and 'Inactive', and tried to classify the status of the occupant by learning environmental data as prediction variables. The major environmental factors affecting the model were identified and the accuracies of prediction according to the classification algorithms were analyzed. As a result, it was confirmed that the main variables influencing the occupant status detection were the lighting electricity energy consumption and CO2 concentration. In addition, we confirmed that by combining these two variables, we can implement an occupant status detection model with a prediction accuracy of 92%. |