Acceleration of the EM algorithm
نویسنده
چکیده
The EM algorithm is used for many applications including 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 “Fisher’s scoring method”. Its convergence is faster, but the calculation 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 show some numerical simulations with our algorithm.
منابع مشابه
A QUASI - NEWTON ACCELERATION OF THE EM ALGORITHM Kenneth
The EM algorithm is one of the most commonly used methods of maximum likelihood estimation. In many practical applications, it converges at a frustratingly slow linear rate. The current paper considers an acceleration of the EM algorithm based on classical quasi-Newton optimization techniques. This acceleration seeks to steer the EM algorithm gradually toward the Newton-Raphson algorithm, which...
متن کاملThe EM Algorithm in Multivariate Gaussian Mixture Models using Anderson Acceleration
Over the years analysts have used the EM algorithm to obtain maximum likelihood estimates from incomplete data for various models. The general algorithm admits several appealing properties such as strong global convergence; however, the rate of convergence is linear which in some cases may be unacceptably slow. This work is primarily concerned with applying Anderson acceleration to the EM algor...
متن کاملAcceleration of the EM algorithm using the vector epsilon algorithm
The Expectation-Maximization (EM) algorithm is a very general and popular iterative computational algorithm to find maximum likelihood estimates from incomplete data and broadly used to statistical analysis with missing data, because of its stability, flexibility and simplicity. However, it is often criticized that the convergence of the EM algorithm is slow. The various algorithms to accelerat...
متن کاملModified frame algorithm and its convergence acceleration by Chebyshev method
The aim of this paper is to improve the convergence rate of frame algorithm based on Richardson iteration and Chebyshev methods. Based on Richardson iteration method, we first square the existing convergence rate of frame algorithm which in turn the number of iterations would be bisected and increased speed of convergence is achieved. Afterward, by using Chebyshev polynomials, we improve this s...
متن کاملRelaxed ordered subsets algorithm for image restoration of confocal microscopy
The expectation−maximization (EM) algorithm for maximum likelihood image recovery converges very slowly. Thus, the ordered subsets EM (OS−EM) algorithm has been widely used in image reconstruction for tomography due to an order−of−magnitude acceleration over the EM algorithm [1]. However, OS− EM is not guaranteed to converge. The recently proposed ordered subsets, separable paraboloidal surroga...
متن کاملOptimum Parameters for Tuned Mass Damper Using Shuffled Complex Evolution (SCE) Algorithm
This study is investigated the optimum parameters for a tuned mass damper (TMD) under the seismic excitation. Shuffled complex evolution (SCE) is a meta-heuristic optimization method which is used to find the optimum damping and tuning frequency ratio for a TMD. The efficiency of the TMD is evaluated by decreasing the structural displacement dynamic magnification factor (DDMF) and acceleration ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
- Systems and Computers in Japan
دوره 31 شماره
صفحات -
تاریخ انتشار 2000