| English | 中文 |
Shared means applicable across stages; it does not establish automatic execution or scientific validity. Stage 02 contains conceptual methods only and does not inherit concrete data or implementation from shared entries.
| Method | Applicable scope | Entry |
|---|---|---|
| Provenance, errors and normalization | Acquisition/preparation and downstream use; stage 02 remains conceptual | Common guidance; stage methods own specialist quality rules |
| Literature | All stages | Literature methods |
| Population and health data | 01 and 04 | Population/health; acquisition skills retain their homes |
| Spatial analysis | Matching, coordinate conversion and mapping across 01, 03 and 04 | Spatial relationships, precision and support; choose the R or Python path in the skill |
| Requirements-driven model design | Defined by the question; stage 02 remains conceptual | Target, validation and reproduction design, pre-experiment planning and review; operational recipes live in workflows |
| Model reproduction and assessment | Reviewed tasks across stages | Comparison and reliability, combined by the model_reproduction workflow |
| Machine learning | Reviewed tasks across stages | Method selection and regression/interpretation; execution belongs to skills |
| Time-series forecasting | Time-ordered prediction with historical external variables | Forecast origin, windows and evaluation; code configuration belongs to the selected forecasting skill |
| Air-pollution regression design | Concentration in 03 and impacts in 04; reference drivers from 01 | Variable roles, validation, causal interpretation and parameter design; operational cases belong to the regression skill |
Stage and SKILL mappings are maintained in index.yaml.
Output analysis, control costs and distributional scenarios are maintained in stage 04 methods; other stages reference the same entries.