Generalising the idea of the classical EM algorithm that is widely used for computing maximum likelihood estimates, we propose an EM-Control (EM-C) algorithm for solving multi-period finite time horizon stochastic control problems. The new algorithm sequentially updates the control policies in each time period using Monte Carlo simulation in a forward-backward manner; in other words, the algori...

We compare three approaches to learning numerical parameters of Bayesian networks from continuous data streams: (1) the EM algorithm applied to all data, (2) the EM algorithm applied to data increments, and (3) the online EM algorithm. Our results show that learning from all data at each step, whenever feasible, leads to the highest parameter accuracy and model classification accuracy. When fac...

This paper advocates a new subspace system identification algorithm for the errorsin-variables (EIV) state space model via the EM algorithm. To initialize the EM algorithm an initial estimate is obtained by the errors-in-variables subspace system identification method: EIV-MOESP (Chou et al. [1997]) and EIV-N4SID (Gustafsson [2001]). The EM algorithm is an algorithm to compute the maximum value...

Clustering is a useful tool for finding structure in a data set. The mixture likelihood approach to clustering is a popular clustering method, in which the EM algorithm is the most used method. However, the EM algorithm for Gaussian mixture models is quite sensitive to initial values and the number of its components needs to be given a priori. To resolve these drawbacks of the EM, we develop a ...

The HMM (Hidden Markov Model) is a probabilistic model of the joint probability of a collection of random variables with both observations and states. The GMM (Gaussian Mixture Model) is a finite mixture probability distribution model. Although the two models have a close relationship, they are always discussed independently and separately. The EM (Expectation-Maximum) algorithm is a general me...

We propose an Expectation-Maximization (EM) algorithm which works on binary decision diagrams (BDDs). The proposed algorithm, BDD-EM algorithm, opens a way to apply BDDs to statistical learning. The BDD-EM algorithm makes it possible to learn probabilities in statistical models described by Boolean formulas, and the time complexity is proportional to the size of BDDs representing them. We apply...

We perform sequence estimation for CPM signals transmitted in a time varying multipath channel. The EM (Expectation-Maximization) algorithm, an iterative proce dure for producing maximum likelihood estimates, is applied to handle the unknown channel. In order to enable implementation of the EM algorithm in this system, a simplification of this algorithm is derived. Channel estimates derived fro...

Introduction My aim is to introduce the Expectation Maximization EM algorithm to you especially some of its theory I will skip proofs but I will derive many formulae that have practical use The EM algorithm is iterative and you should be familiar with its convergence properties I will discuss them in detail I will present applications of the EM algorithm to signal and image processing in a comp...

In a previous class, we discussed an algorithm for learning a probabilistic matrix model which describes a fixed-length motif in a set of sequences S : : : S over an alphabet A. This algorithm is one of a class of methods collectively known as expectation maximization, or EM. We will describe the general EM algorithm, then derive the motif-finding algorithm by applying EM to learn a specific pr...