Precipitation Forecasting Using Transformer: A Comparative Study with Unet
DOI:
https://doi.org/10.54097/hset.v39i.6613Keywords:
Transformer; Unet; Precipitation Forecast.Abstract
Accurate precipitation prediction has a huge socio-economic impact. With the development of algorithms, architectures, and hardware technologies in the field of machine learning, the disadvantages of traditional numerical weather prediction methods are becoming more and more obvious. Noting that related studies have attempted to use the latest machine learning architectures for precipitation prediction, this work uses Unet and Transformer for precipitation prediction based on cloud layer information and compares them. The results indicate that Transformer can achieve 94.13% accuracy, while Unet has 96.89%. Finally, in conducting the data comparison, a conjecture related to the lagging of Transformer data is proposed based on the results of this experiment.
Downloads
References
Lorenc A C. Analysis methods for numerical weather prediction[J]. Quarterly Journal of the Royal Meteorological Society, 1986, 112(474): 1177-1194.
Schultz M G, Betancourt C, Gong B, et al. Can deep learning beat numerical weather prediction? [J]. Philosophical Transactions of the Royal Society A, 2021, 379(2194): 20200097.
Sønderby C K, Espeholt L, Heek J, et al. Metnet: A neural weather model for precipitation forecasting[J]. arXiv preprint arXiv:2003.12140, 2020.
Han K, Xiao A, Wu E, et al. Transformer in Transformer[J]. Advances in Neural Information Processing Systems, 2021, 34: 15908-15919.
Jin W, Barzilay R, Jaakkola T. International conference on machine learning[J]. 2020.
Huang H, Lin L, Tong R, et al. 3+: A Full-Scale Connected UNet for Medical Image Segmentation. arXiv 2020[J]. arXiv preprint arXiv:2004.08790.
Li X, Chen H, Qi X, et al. H-DenseUNet: hybrid densely connected UNet for liver and tumor segmentation from CT volumes[J]. IEEE transactions on medical imaging, 2018, 37(12): 2663-2674.
Trebing K, Staǹczyk T, Mehrkanoon S. SmaAt-UNet: Precipitation nowcasting using a small attention-UNet architecture[J]. Pattern Recognition Letters, 2021, 145: 178-186.
Cao H, Wang Y, Chen J, et al. Swin-unet: Unet-like pure Transformer for medical image segmentation[J]. arXiv preprint arXiv:2105.05537, 2021.
Wang H, Xie S, Lin L, et al. Mixed Transformer u-net for medical image segmentation[C]//ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP). IEEE, 2022: 2390-2394.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
