3.4. Convergence Problems
As explained in the previous section, maximum likelihood estimation of the logit model is an iterative process of successive approximations. When the change in the coefficient estimates from one iteration to the next is very small, the computations stop and the algorithm is said to have converged. Usually this process goes smoothly with no special attention needed. However, sometimes the iterative process breaks down so that convergence is not achieved. Dealing with convergence failures can be one of the more frustrating problems encountered by users of logit regression.
LOGISTIC has a default limit of 25 iterations and GENMOD has a default limit of 50. If the algorithm hasn’t converged by this limit, both procedures ...
Get Logistic Regression Using SAS®: Theory and Application now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
