Properties2
| Type | Concept |
| Note created | Feb 17, 2025 |
Collinearity is a linear relationship that is present between two explanatory variables. If the correlation between them is 1 or -1, it is considered that these two variables are perfectly collinear. Despite it implies the existence of a linear relationship, it does not refer to the causality or lack thereof between the two variables.
Collinearity is an undesired property in regression models, since it can lead to skewed or noisy results. Ideally, only one of the two (or more, if there is multicollinearity present) variables should be chosen to be used as explanatory.