| 作成者 |
|
|
|
|
|
|
|
|
|
|
|
| 本文言語 |
|
| 出版者 |
|
| 発行日 |
|
| 収録物名 |
|
| 巻 |
|
| 開始ページ |
|
| 終了ページ |
|
| 会議情報 |
|
| 出版タイプ |
|
| アクセス権 |
|
| Crossref DOI |
|
| 権利関係 |
|
| 権利関係 |
|
| 概要 |
Predicting indoor absolute humidity is essential for effective building management, energy efficiency and occupant comfort. This research evaluates the performance of Transformer and Informer models i...n predicting indoor humidity using outdoor humidity data. The dataset consists of outdoor humidity measurements to forecast indoor conditions. We developed Transformer and Informer models to focus on attention mechanisms and sequence generation. Performance was evaluated using metrics such as MAPE, MSE, MAE, RMSE and R-squared. The Transformer model slightly outperforms the Informer model with a MAPE of 3.42% and an R-squared of 0.911, compared to the Informer's MAPE of 4.55% and R-squared of 0.867. This superior performance is due to the Transformer's enhanced attention mechanism and efficient sequence handling. This study provides advanced models for accurate indoor humidity prediction, with significant implications for building management and energy savings. Future research could explore real-time implementation and application to other environmental parameters.続きを見る
|
Mendeley出力
2024_p1084