건물에너지 소비량 예측을 위한 기계학습방법 선정연구
정진화, 채영태 · 대한설비공학회 2017년도 하계학술발표대회 · 2017
Building Energy Consumption Forecasting is very important in demand management in buildings. For this reason, many studies on Building Energy Consumption Forecasting are proceeding. But forecasting accuracy is decreased because of data processing, suitability of machine learning types. In this study, forecasting accuracy of various machine learning types based on weather data was compared and analyzed to select machine learning type. By using Artificial Neural Network (ANN), Support Vector Machine (SVM) and Random Forest (RF), we compared and analyzed using performance evaluation indicator such as Mean Bias Error(MBE), Coefficient of Variation of the Root Mean Square Error (CV(RMSE)). The results demonstrate that RF showed best performance followed ANN, SVM.
Cited by 0 works.