Maximum Likelihood Approach
Arijit Chaudhuri · Statistics, a series of textbooks and monographs · 2010
E I py p y ( ) ( )( ) ( ) ( ) , = + − − = − + − 1 1 1 2 1 leading to r I p pi i= − − −[ ( )]/( )1 2 1 and ER(ri) = yi. Again, since I Ii i2 = and y yi i2 = , V I E I E I p p V r p p ( ) ( )( ( )) ( ), ( ) ( ) = − = − = = − 1 1 1 and hence Vi ( ) .