Solution
ID: past-exam-of-the-mathematics-course-of-the-university-of-cambridge/2009/iii/paper-40/6/a/solution
Past exam of the mathematics course of the University of Cambridge 2009 iii Paper 40 6 a Solution by
Codex 0 Created 2026-10-03 Updated 2026-10-14
The observed-data likelihood function integrates out the latent variable:Choose an initial parameter with positive likelihood. At iteration , the E-step of the EM algorithm forms the entire expected complete-data log-likelihoodThe expectation is under the conditional missing-data law at the old parameter, while the argument remains free. The M-step then choosesThe observed data stay fixed throughout. This is not obtained merely by substituting the conditional mean of into a nonlinear log-likelihood: the expectation of all required complete-data terms must be taken. A generalized EM step only requires , which suffices for the monotonicity proved next. Exact maximization assumes that an admissible maximizer exists.
New to topics? Read the docs here!