Development of a Hybrid Attention Transformer for Daily PM<sub>2.5</sub> Predictions in Seoul
A hybrid attention transformer (HAT) was developed for accurate daily PM<sub>2.5</sub> predictions in Seoul. The performance of the HAT was evaluated through a comparative analysis of its predictions against ground-based observations and those from a three-dimensional chemical transport...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-01-01
|
| Series: | Atmosphere |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-4433/16/1/37 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832589149549363200 |
|---|---|
| author | Hyun S. Kim Kyung M. Han Jinhyeok Yu Nara Youn Taehoo Choi |
| author_facet | Hyun S. Kim Kyung M. Han Jinhyeok Yu Nara Youn Taehoo Choi |
| author_sort | Hyun S. Kim |
| collection | DOAJ |
| description | A hybrid attention transformer (HAT) was developed for accurate daily PM<sub>2.5</sub> predictions in Seoul. The performance of the HAT was evaluated through a comparative analysis of its predictions against ground-based observations and those from a three-dimensional chemical transport model (3-D CTM). The results demonstrated that the HAT outperformed the 3-D CTM, achieving a 4.60% higher index of agreement (IOA). Additionally, the HAT exhibited 22.09% fewer errors and 82.59% lower bias compared to the 3-D CTM. Diurnal variations in PM<sub>2.5</sub> predictions from both models were also analyzed to explore the characteristics of the proposed model further. The HAT predictions closely aligned with observed PM<sub>2.5</sub> throughout the day, whereas the 3-D CTM exhibited significant diurnal variability. The importance of the input features was evaluated using the permutation method, which revealed that the previous day’s PM<sub>2.5</sub> was the most influential feature. The robustness of the HAT was further validated through a comparison with the long short-term memory (LSTM) model, which showed 18.50% lower errors and 95.91% smaller biases, even during El Niño events. These promising findings highlight the significant potential of the HAT as a cost-effective and highly accurate tool for air quality prediction. |
| format | Article |
| id | doaj-art-e033ba9e6ede4c1aa08e3dcb9ec00396 |
| institution | Kabale University |
| issn | 2073-4433 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Atmosphere |
| spelling | doaj-art-e033ba9e6ede4c1aa08e3dcb9ec003962025-01-24T13:21:47ZengMDPI AGAtmosphere2073-44332025-01-011613710.3390/atmos16010037Development of a Hybrid Attention Transformer for Daily PM<sub>2.5</sub> Predictions in SeoulHyun S. Kim0Kyung M. Han1Jinhyeok Yu2Nara Youn3Taehoo Choi4School of Environment and Energy Engineering, Gwangju Institute of Science and Technology (GIST), Gwangju 61005, Republic of KoreaSchool of Environment and Energy Engineering, Gwangju Institute of Science and Technology (GIST), Gwangju 61005, Republic of KoreaSchool of Environment and Energy Engineering, Gwangju Institute of Science and Technology (GIST), Gwangju 61005, Republic of KoreaSchool of Environment and Energy Engineering, Gwangju Institute of Science and Technology (GIST), Gwangju 61005, Republic of KoreaSchool of Environment and Energy Engineering, Gwangju Institute of Science and Technology (GIST), Gwangju 61005, Republic of KoreaA hybrid attention transformer (HAT) was developed for accurate daily PM<sub>2.5</sub> predictions in Seoul. The performance of the HAT was evaluated through a comparative analysis of its predictions against ground-based observations and those from a three-dimensional chemical transport model (3-D CTM). The results demonstrated that the HAT outperformed the 3-D CTM, achieving a 4.60% higher index of agreement (IOA). Additionally, the HAT exhibited 22.09% fewer errors and 82.59% lower bias compared to the 3-D CTM. Diurnal variations in PM<sub>2.5</sub> predictions from both models were also analyzed to explore the characteristics of the proposed model further. The HAT predictions closely aligned with observed PM<sub>2.5</sub> throughout the day, whereas the 3-D CTM exhibited significant diurnal variability. The importance of the input features was evaluated using the permutation method, which revealed that the previous day’s PM<sub>2.5</sub> was the most influential feature. The robustness of the HAT was further validated through a comparison with the long short-term memory (LSTM) model, which showed 18.50% lower errors and 95.91% smaller biases, even during El Niño events. These promising findings highlight the significant potential of the HAT as a cost-effective and highly accurate tool for air quality prediction.https://www.mdpi.com/2073-4433/16/1/37artificial neural networkhybrid attention transformerdaily PM<sub>2.5</sub> prediction |
| spellingShingle | Hyun S. Kim Kyung M. Han Jinhyeok Yu Nara Youn Taehoo Choi Development of a Hybrid Attention Transformer for Daily PM<sub>2.5</sub> Predictions in Seoul Atmosphere artificial neural network hybrid attention transformer daily PM<sub>2.5</sub> prediction |
| title | Development of a Hybrid Attention Transformer for Daily PM<sub>2.5</sub> Predictions in Seoul |
| title_full | Development of a Hybrid Attention Transformer for Daily PM<sub>2.5</sub> Predictions in Seoul |
| title_fullStr | Development of a Hybrid Attention Transformer for Daily PM<sub>2.5</sub> Predictions in Seoul |
| title_full_unstemmed | Development of a Hybrid Attention Transformer for Daily PM<sub>2.5</sub> Predictions in Seoul |
| title_short | Development of a Hybrid Attention Transformer for Daily PM<sub>2.5</sub> Predictions in Seoul |
| title_sort | development of a hybrid attention transformer for daily pm sub 2 5 sub predictions in seoul |
| topic | artificial neural network hybrid attention transformer daily PM<sub>2.5</sub> prediction |
| url | https://www.mdpi.com/2073-4433/16/1/37 |
| work_keys_str_mv | AT hyunskim developmentofahybridattentiontransformerfordailypmsub25subpredictionsinseoul AT kyungmhan developmentofahybridattentiontransformerfordailypmsub25subpredictionsinseoul AT jinhyeokyu developmentofahybridattentiontransformerfordailypmsub25subpredictionsinseoul AT narayoun developmentofahybridattentiontransformerfordailypmsub25subpredictionsinseoul AT taehoochoi developmentofahybridattentiontransformerfordailypmsub25subpredictionsinseoul |