Temperature compensation or a real structural change?
Correlation between a measurement and temperature is not enough to subtract temperature effects. See the procedure from raw data through the model and residual to engineering validation.
- What temperature compensation really means
- Three questions before the first regression
- Procedure: raw data + temperature -> model -> residual -> validation
- How to read the results without confusing correlation with cause
- Illustrative example: a model that helps, but has no right to switch off vigilance
- How to set alarms with compensation
- Warning signs: when to suspend compensation
- What this looks like in Inclify
Read article
13 min read