Analyse factorielle



Définition

L'analyse factorielle est une méthode statistique utilisée pour décrire la variabilité entre les variables corrélées observées en termes d'un nombre potentiellement inférieur de variables non observées appelées facteurs. Par exemple, il est possible que les variations de six variables observées reflètent principalement les variations de deux variables (sous-jacentes) non observées.

Français

analyse factorielle loc. nom. fém.

Anglais

Factor analysis


Factor analysis is a method for modeling observed variables and their covariance structure in terms of unobserved variables (i.e., factors). There are two types of factor analyses, exploratory and confirmatory. Exploratory factor analysis (EFA) is method to explore the underlying structure of a set of observed variables, and is a crucial step in the scale development process. The first step in EFA is factor extraction. Common factor analysis models can be estimated using various estimation methods such as principal axis factoring and maximum likelihood, and we will compare the practical differences between these two methods. After extracting the best factor structure, we can obtain a more interpretable factor solution through factor rotation. Confirmatory factor analysis (CFA), which is a method to verify a factor structure that has already been defined. Topics to discuss include identification, model fit, and examples of a one-factor, uncorrelated two-factor and correlated two-factor model.

Source: https://stats.idre.ucla.edu/spss/seminars/introduction-to-factor-analysis/


Source: wikipedia, Analyse factorielle.

Source: L'apprentissage profond, Ian Goodfellow, Yoshua Bengio et Aaron Courville Éd. Massot 2018 page 488