Two Methods for Estimating a Markov Transition Matrix from Subsampled Data
Samantha Massengill · 2011
Professor Bhaskar Krishnamachari 1. In this work, we use adaptive algorithms to estimate 2-state Markovian transition probabilities from a subsampled sequence when these probabilities are not known a priori. A maximum likelihood estimation (MLE) problem is set up, and solved using two different methods: the expectation maximization (EM) algorithm, and an exhaustive search through