EM مثال هايي از الگوريتم

Exercise in EM

Suppose survival times follows an exponential distribution and some observations are right-censored: in this situation the EM algorithm gives strightforward solution to the problem of maximum likelihood estimation. But what happens if survival times are also left - censored, or if they follow a uniform distribution? The EM algorithm is a generic device useful in variety of problems with incomplete data, and it appears more and more often in statistical textbooks. This article presents two exercises, which are extentions of well - known example used in introductions to the EM algorithm. They focus on two ponits: the appliicability of the algorithm and its self-consistency property.