期刊
IEEE ACCESS
卷 9, 期 -, 页码 15413-15425出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/ACCESS.2021.3053317
关键词
Deep learning; individualized models; meta learning; transfer learning; short-term load forecasting
资金
- Electronics and Telecommunications Research Institute (ETRI) - Korean government [20ZR1100]
- Korea Medical Device Development Fund' - Korea government (the Ministry of Science and ICT) [202013B14]
- Korea Medical Device Development Fund' - Korea government (Ministry of Trade, Industry and Energy) [202013B14]
- Korea Medical Device Development Fund' - Korea government (Ministry of Health Welfare, Republic of Korea) [202013B14]
- Korea Medical Device Development Fund' - Korea government (Ministry of Food and Drug Safety) [202013B14]
It is generally believed that individualized models are the best way to predict electric load, but traditional methods tend to favor one-for-all models. This study utilizes transfer learning and meta learning, successfully integrated into deep neural networks, to form high-performance individualized models using individual data in just a few days.
While the general belief is that the best way to predict electric load is through individualized models, the existing studies have focused on one-for-all models because the individual models are difficult to train and require a significantly larger data accumulation time per individual. In recent years, applying deep learning for forecasting electric load has become an important research topic but still one-for-all has been the main approach. In this work, we adopt transfer learning and meta learning that can be smoothly integrated into deep neural networks, and show how a high-performance individualized model can be formed using the individual's data collected over just several days. This is made possible by extracting the common patterns of many individuals using a sufficiently large dataset, and then customizing each individual model using the specific individual's small dataset. The proposed methods are evaluated over residential and non-residential datasets. When compared to the conventional methods, the meta learning model shows 7.84% and 15.07% RMSE improvements over the residential and non-residential datasets, respectively. Our results suggest that the individualized models can be used as effective tools for many short-term load forecasting tasks.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据