- are not constant and
- changes as function of the regression independent variables.
The squared residual (i.e, residual is the estimated error) is used to proxy the error variance.
Monday, May 10, 2010
Conditional Heteroskedasticity
A regression exhibits conditional heteroskedasticity if the variance of the regression errors:
The squared residual (i.e, residual is the estimated error) is used to proxy the error variance.
The squared residual (i.e, residual is the estimated error) is used to proxy the error variance.
Subscribe to:
Post Comments (Atom)
0 comments:
Post a Comment