Acceleration of the EM algorithm using the Vector Aitken method and its Steffensen form (Q523777)

The authors propose an acceleration expectation-maximization (EM) and vector Aitken (VA) accelerated EM algorithms, whose convergence speed is faster than that of the EM algorithm. The algorithms are based on the use of the VA method. The VA-accelerated EM algorithm does not use the information matrix but only uses the sequence of estimates obtained from iterations of the EM algorithm, thus it keeps the flexibility and simplicity of the EM algorithm. Th. Steffensen's iterative process is also considered and the Steffensen form of the VA-accelerated EM algorithm is provided. It is shown that the reform process has a quadratic convergence. Some numerical examples are presented to illustrate the efficiency of the proposed methods.