How can the covariance of 1xn score function be computed for fisher information

108 Views Asked by At

Fisher Information can be calculated using the covariance of score function. For logistic regression the log-likelihood is taken to obtain the score function which is

 YX - PX              // Y-class X-record P-htheta

On computing the result of this I get a 1xn matrix. And I am confused how to calculate the covariance of this score function. I know how generally a covariance matrix is calculated but I am stuck with this 1xn. Please help.