一般化された線形モデルフィット関数のオプションで,パラメータ共分散行列の推定法を指定する.
CovarianceEstimatorFunction
一般化された線形モデルフィット関数のオプションで,パラメータ共分散行列の推定法を指定する.
詳細
- CovarianceEstimatorFunctionは,GeneralizedLinearModelFit,LogitModelFit,ProbitModelFitのオプションである.
- 可能な設定値には,予測情報行列を使う"ExpectedInformation"と観察情報行列を使う"ObservedInformation"がある.
- 共分散行列は
に等しい.ただし,ϕ は分散パラメータ,
はフィッシャーの情報行列である.
例題
すべて開く すべて閉じる例 (2)
予測される情報行列を使って一般化された線形モデルをフィットする:
data = {{0, 1}, {1, 1.5}, {3, 2}, {5, 10}};glm = GeneralizedLinearModelFit[data, x, x, ExponentialFamily -> "Gamma", CovarianceEstimatorFunction -> "ExpectedInformation"]glm["CovarianceMatrix"]data = {{0, 1}, {1, 1.5}, {3, 2}, {5, 10}};glm = GeneralizedLinearModelFit[data, x, x, ExponentialFamily -> "Gamma"]glm["CovarianceMatrix", CovarianceEstimatorFunction -> "ObservedInformation"]glm["CovarianceMatrix"]スコープ (3)
FittedModel内で共分散の推定を指定する:
data = {{0, 1}, {1, 1.5}, {3, 2}, {5, 10}};glm = GeneralizedLinearModelFit[data, x, x, ExponentialFamily -> "Gamma"]glm["CovarianceMatrix", CovarianceEstimatorFunction -> "ObservedInformation"]LogitModelFitとともに使う:
data = {.15, .27, .58, .49, .77};logit = LogitModelFit[data, x, x]logit["ParameterErrors"]logit["ParameterErrors", CovarianceEstimatorFunction -> "ObservedInformation"]ProbitModelFitとともに使う:
data = {.15, .27, .58, .49, .77};probit = ProbitModelFit[data, x, x]probit["ParameterErrors"]probit["ParameterErrors", CovarianceEstimatorFunction -> "ObservedInformation"]特性と関係 (2)
data = {{0, 1}, {1, 1.5}, {3, 2}, {5, 10}};glm = GeneralizedLinearModelFit[data, x, x, ExponentialFamily -> "Gamma"]glm[{"ParameterErrors", "ParameterConfidenceIntervals"}]glm[{"ParameterErrors", "ParameterConfidenceIntervals"}, CovarianceEstimatorFunction -> "ObservedInformation"]CovarianceEstimatorFunctionは共分散の一般的な構造を制御する:
glm = GeneralizedLinearModelFit[{{0, 1}, {1, 1.5}, {3, 2}, {5, 10}}, x, x, ExponentialFamily -> "Gamma"]{err, disp} = glm[{"ParameterErrors", "EstimatedDispersion"}, CovarianceEstimatorFunction -> "ObservedInformation"]DispersionEstimatorFunctionはスケールに影響する:
{err2, disp2} = glm[{"ParameterErrors", "EstimatedDispersion"}, CovarianceEstimatorFunction -> "ObservedInformation", DispersionEstimatorFunction -> (1&)](err / err2) ^ 2disp / disp2テキスト
Wolfram Research (2008), CovarianceEstimatorFunction, Wolfram言語関数, https://reference.wolfram.com/language/ref/CovarianceEstimatorFunction.html.
CMS
Wolfram Language. 2008. "CovarianceEstimatorFunction." Wolfram Language & System Documentation Center. Wolfram Research. https://reference.wolfram.com/language/ref/CovarianceEstimatorFunction.html.
APA
Wolfram Language. (2008). CovarianceEstimatorFunction. Wolfram Language & System Documentation Center. Retrieved from https://reference.wolfram.com/language/ref/CovarianceEstimatorFunction.html
BibTeX
@misc{reference.wolfram_2026_covarianceestimatorfunction, author="Wolfram Research", title="{CovarianceEstimatorFunction}", year="2008", howpublished="\url{https://reference.wolfram.com/language/ref/CovarianceEstimatorFunction.html}", note=[Accessed: 15-September-2026]}
BibLaTeX
@online{reference.wolfram_2026_covarianceestimatorfunction, organization={Wolfram Research}, title={CovarianceEstimatorFunction}, year={2008}, url={https://reference.wolfram.com/language/ref/CovarianceEstimatorFunction.html}, note=[Accessed: 15-September-2026]}