Acceleration of the EM algorithm
Shiro Ikeda · Systems and Computers in Japan · 2000
The EM algorithm is used for many applications, including the Boltzmann machine, stochastic Perceptron, and HMM. This algorithm gives an iterating procedure for calculating the MLE of stochastic models which have hidden random variables. It is simple, but the convergence is slow. We also have the “Fisher scoring method.” Its convergence is faster, but the calculation load is heavy. We show that by using the EM algorithm recursively, we can connect these two methods and accelerate the EM algorithm. Also, Louis, Meng, and Rubin showed they can accelerate the EM algorithm, but our algorithm is simpler. We present some numerical simulations using our algorithm. © 2000 Scripta Technica, Syst Comp Jpn, 31(2): 10–18, 2000