How can I characterize (bias, efficiency) the ratio of two discrete likelihood estimators qˆ = ˆN1 (x)/ ˆN(x) where ˆN(x )= ˆN1(x )+ ˆN2(x)?

Where ^N1(x)=Sum(1,xj=1), j=1,N and xj=1 for j in N1, 0 otherwise

21 May 2014 5,596 1 View