Порівняння методів

	Робастний факторний аналіз	Діагностика впливу (відстань Кука, DFFITS, плече)	Метод головних компонент
Галузь	Статистика	Статистика	Машинне навчання
Родина	Regression model	Regression model	Machine learning
Рік появи	2003	1977	2002
Автор методу	Pison, Rousseeuw, Filzmoser & Croux	R. Dennis Cook (Cook's distance); Belsley, Kuh & Welsch (DFFITS, leverage)	Jolliffe, I.T. (textbook); Pearson & Hotelling (origins)
Тип	Robust latent-factor model	Regression diagnostic	Unsupervised dimensionality reduction
Основоположне джерело	Pison, G., Rousseeuw, P. J., Filzmoser, P., & Croux, C. (2003). Robust factor analysis. Journal of Multivariate Analysis, 84(1), 145-172.	Cook, R. D. (1977). Detection of Influential Observations in Linear Regression. Technometrics, 19(1), 15-18.	Jolliffe, I.T. (2002). Principal Component Analysis (2nd ed.). Springer.
Інші назви	robust factor analysis, outlier-resistant factor analysis, MCD-based factor analysis, Robust Faktör Analizi	Cook's distance, DFFITS, leverage, influential observation detection	Temel Bileşenler Analizi (PCA), PCA, principal components analysis, Karhunen-Loève transform
Пов'язані	5	5	3
Підсумок	Robust Factor Analysis recovers the latent factor structure of multivariate continuous data while resisting the distorting pull of outliers. Introduced by Pison, Rousseeuw, Filzmoser and Croux (2003), it replaces the classical sample covariance with a robust estimator such as the Minimum Covariance Determinant (MCD) or an S-estimator before extracting factors.	Influence diagnostics are a family of post-fit measures that quantify how much each single observation affects a fitted regression. Cook's distance was introduced by R. Dennis Cook in 1977, with leverage and DFFITS formalised by Belsley, Kuh and Welsch in 1980, to flag the observations that most strongly pull the estimated coefficients.	Principal Component Analysis (PCA) is an unsupervised dimensionality-reduction method — given its modern textbook treatment by Ian Jolliffe (2002) — that compresses high-dimensional data into fewer dimensions while preserving the maximum possible variance. It re-expresses correlated variables as a small set of uncorrelated principal components ordered by how much of the data's variation each one captures.