- Python 90.6%
- R 9.4%
| Filename | Latest commit message | Latest commit date |
|---|---|---|
| Evaluation_pipeline | ||
| level1_data | ||
| meteoswiss_2022-operational | ||
| meteoswiss_2025-Beta-1 | ||
| meteoswiss_2025-Beta-2 | ||
| meteoswiss_2025-Gamma-1 | ||
| meteoswiss_2025-Gamma-3 | ||
| meteoswiss_2025-Omega-1 | ||
| meteoswiss_2025-Omega-2 | ||
| meteoswiss_2025_Q2_15sp | ||
| .gitignore | ||
| LICENSE | ||
| README.md | ||
| SI_1_training datasets.csv | ||
| SI_2_metrics.csv | ||
| SI_3_metrics_area.csv | ||
| SI_4_corr_pred_ref.csv | ||
| SI_5_flow.csv | ||
| SI_6_2022Operational.csv | ||
| SI_6_P33_2025Gamma2.csv | ||
The meteoswiss_2025_Q2_15sp (2025-Gamma-2) model was developed and trained for use within the Swiss Pollen Monitoring Network. It has been tested and validated exclusively under Swiss conditions for seven pollen taxa: Alnus, Betula, Corylus, Fagus, Fraxinus, Poaceae, and Quercus. The model is described in a scientific publication:
Operational pollen classification using digital holography and fluorescence and labeled "2025-Gamma-2" in the preprint "A general framework for evaluating real-time bioaerosol classification algorithms https://doi.org/10.5194/egusphere-2025-5440".
Although the model technically includes a total of 15 classes, only the seven taxa listed above have been validated. The remaining classes have not been validated and may produce unreliable or inaccurate results. Use of the model outside its intended scope—including applications in other geographic regions, for unvalidated taxa, or for operational decision-making—is not recomended.
The ONNX file contains the model. The text file lists the thresholds and classes used operationally. The xlsx file contains description of and links to training datasets. Model architecture is illustrated in the png file produced by Netron.app Version 8.4.4.
Alternative models tested by MeteoSwiss are provided for reference as well.