I´m trying do slove a hedonic GWR model for house prices but I have some extrem coefients 😞 !
As I have found out I had to do something with the collinearity of the variables and that there are several "types" of collinearity. So you can group them in:
- global collinearity and - local collinearity
1. Question: Is collinearity and multicollinearity the same/wahts the difference? (Because in both context global/local they are used and I have found no easy explanation in words)
2. Question: I have checked: + global collinearity with [INDENT]- global "multi"collinearity (tested with VIF (variance inflation factor) and ok - all values < 3,5)[/INDENT] + local collinearity with [INDENT]- local "multi"collinearity (tested with variation of variables in ArcGIS GWR - all condition number < 30)[/INDENT] But I still have tow variables in the model with very extrem coefients which indicates collinearity as I think!
So I have do to another collinearity or mulitcollinearity check?