A new class of statistical algorithms is presented and examined. The method is called the α-EM algorithm. This novel algorithm contains the traditional EM algorithm as a special case of α = -1. The choice of the design parameter `α' affects the eigenvalues of Hessian matrices for likelihood maximization. This causes much faster convergence than the traditional EM algorithm. Convergence theorems are given for the basic α-EM algorithm and its practical variants. Numerical evaluation shows fast convergence at nearly one-third the iteration counts and one-half the CPU time relative to the traditional method.
|ジャーナル||Systems and Computers in Japan|
|出版ステータス||Published - 2000 10|
ASJC Scopus subject areas
- Computational Theory and Mathematics
- Hardware and Architecture
- Information Systems
- Theoretical Computer Science