Abstract
| - Equation (6) represents only an approximation to the maximum likelihood estimator, whose actual form satisfies π^(t)=∑i.yi∈Iwiexp(−β^biai)ai/∑i:yi∈Iwiexp(−β^biai) where wiα[{1−π(t)}exp(−β^bi)+π(t)]−1 Our reason for presenting (6) is its generalizability to the closed form, consistent prevalence estimator, (8).
|