α-EM algorithm and its basic properties

Yasuo Matsuyama

    研究成果: Article査読

    抄録

    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.

    本文言語English
    ページ(範囲)12-23
    ページ数12
    ジャーナルSystems and Computers in Japan
    31
    11
    DOI
    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

    フィンガープリント 「α-EM algorithm and its basic properties」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

    引用スタイル