Tutorials : influential points
> Influential points
Who to detect?
In linear or logistic regression, Cook's distances are commonly used to detect influential values (when the absolute value is greater than 1).
What are the consequences?
The robustness of the model can be questioned. Indeed, if these individuals were not observed, the results would differ considerably.
You can check the data source for these individuals to ensure that there are no errors.
How to manage?