| |
|
Science Forum Index » Statistics - Math Forum » inverse of the information matrix. var-covar matrix of param
Page 1 of 1
|
| Author |
Message |
| quebecstat |
 Posted: Fri Jan 26, 2007 6:40 pm |
|
|
|
Guest
|
Fact:
The estimated large sample var-covar matrix for maximum likelihood estimators is the inverse of the information matrix. The elements of the information matrix are the negatives of the expected values of the second-order partial derivatives of the log of the likelihood function evaluated at the value of the parameter.
Question:
Show that for LOGISTIC REGRESSION, the var-covar matrix is the inverse of a different (more simple) matrix, the elements of which are the negatives of the second-order partial derivatives of the log of the likelihood function evaluated at the value of the parameter.
in other words, the expectation sign "cancels out" for the logistic regression.
any idea/pointer? |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Fri Oct 10, 2008 8:16 pm
|
|
