Model Based Inference in the Life Sciences: A Primer on Evidence David R. Anderson Model Based Inference in the L

CHAPTER 2 Question 2. One worry might be that the algorithm did not converge to the maximum. This could be because the log-likelihood was very flat near the maximum point or that sparce data cause a lack of estimability. A second concern might be that mixture distributions are not in the exponential family and the log-likelihood function might have more than a single mode. Both issues might be helped by using different starting positions for the parameters. A more sophisticated approach would be to use simulated annealing.