| Title |
A Study on Improving Convenience Store Power Consumption Prediction Through the Introduction of Apparent Temperature and Degree Day Variables |
| Authors |
김기한(Ki-Han Kim) ; 김정욱(Jeong-Uk Kim) |
| DOI |
https://doi.org/10.5370/KIEE.2024.73.6.939 |
| Keywords |
Building energy prediction; Accuracy improvement; Ensemble model; Machine learning |
| Abstract |
Accurate prediction is crucial for establishing energy management strategies. This study proposes a method to enhance the accuracy of predicting power consumption in convenience stores by considering the necessity of heating and cooling. We collected building energy usage data and meteorological data from 86 convenience stores nationwide, incorporating perceived temperature and daily variables into machine learning models. Case studies before and after introducing these variables were conducted to evaluate the model's performance, confirming notably improved accuracy, especially with CatBoost and Stacking models. Through this approach, maximizing energy management efficiency and optimizing energy consumption can be achieved in existing buildings without additional equipment installation. |