To optimize a model, follow these steps:
- Use variables or KPIs as output parameters and input parameters for which data is available without gaps in the calculation time range.
- If necessary, remove input parameters with insufficient data quality.
- Select a calculation period during which the data quality is better.
- Change the spike correction to mitigate the effect of extreme variable values.
- Recalculate the formula after each change.
- When you have achieved an acceptable result, save the model.