BROWSE
Detail
Conference Paper
A Study on ML Model Performance Analysis According to Multiple Levels of Learning Data for Building Energy Consumption Prediction
Cited - time in 
Share 
Share
- Authors
- Taehyung Kim, Youn Kwae Jeong, Seok Jin Lee
- Issue Date
- 2021-12
- Citation
- International Conference on Ubiquitous Information Technologies and Applications (CUTE) 2021, pp.1-6
- Language
- English
- Type
- Conference Paper
- Abstract
- Recently, the need for energy saving is increasing all over the world. Therefore, even in the building sector, which consumes a large proportion of energy, research and development for energy saving is being actively conducted. In order to save energy in a building, physical savings through building remodeling or facility replacement are important, but it is also important to induce efficient energy consumption in terms of building operation. In order to use energy efficiently, accurate energy consumption pattern analysis and consumption prediction technology are required. For this purpose, machine learning models have been widely used recently. To develop a machine learning-based building energy consumption prediction model, learning data is required, and the prediction accuracy differs depending on the levels of learning data such as data size, parameters types, input/output steps, etc. In this study, we analyze the correlation between the predictive model performance and the multiple levels of learning data.
- KSP Keywords
- Building sector, Consumption pattern, Data size, Efficient energy consumption, Energy saving, Learning data, Learning-based, Model performance, Multiple levels, Performance analysis, Prediction accuracy