A consistent estimate of the transitional probability matrix of a finite Markov process in the case when at each point in time only the proportions of the sample in each state are known. It is shown that this estimate is asymptotically more efficient...
