Estimation Based on Estimating Equations. From the Maximum Likelihood Method and the Method of Moments.
Takemi Yanagimoto
Abstract
Open-access reader
Takemi Yanagimoto
Abstract
Open-access reader
従来モーメント法は簡便で受け入れ易く,最尤法は計算は困難だが理論的には望ましいと漠然と信じられてきた.今日では複雑なデータを良く記述するためにモデルに含まれる母数の次元が高くなり,より精密な推論が求められている.そこでは最尤推定量は望ましくないので,モーメント法のもつ頑健さを取り入れた推定量の構成が必要になる.実際的には簡単な構造をしたモデル,及びその形式的な一般化,が用いられる.本稿ではこれらの視点からモーメント法と最尤法をレビューする.その上で近年注目を浴びている一般化線形モデル,GEE,推定方程式の感受性規準を系統的に概説する.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
従来モーメント法は簡便で受け入れ易く,最尤法は計算は困難だが理論的には望ましいと漠然と信じられてきた.今日では複雑なデータを良く記述するためにモデルに含まれる母数の次元が高くなり,より精密な推論が求められている.そこでは最尤推定量は望ましくないので,モーメント法のもつ頑健さを取り入れた推定量の構成が必要になる.実際的には簡単な構造をしたモデル,及びその形式的な一般化,が用いられる.本稿ではこれらの視点からモーメント法と最尤法をレビューする.その上で近年注目を浴びている一般化線形モデル,GEE,推定方程式の感受性規準を系統的に概説する.
Key concepts: Mathematics, Estimating equations, Maximum likelihood, Generalized method of moments, Gee, Method of moments (probability theory), Generalized estimating equation, Statistics