scientific article; zbMATH DE number 3999030

From MaRDI portal
Publication:4725536





In this note we have a random sample of observations \(x_ 1,x_ 2,...,x_ n\) from a multivariate distribution with mean \(\mu =(\mu_ 1,...,\mu_ p)'\) and covariance matrix \(\Sigma =(\sigma_{ij})\). In the orthogonal factor analysis model, the covariance matrix is \(\Sigma =\Lambda \Lambda '+\Psi\), where \(\Lambda =(\lambda_{ij})\) is the factor loadings matrix and \(\Psi =(\psi_{ij})\) is the covariance of the specific factor and \(\Psi\) is diagonal matrix. The first problem of factor analysis is to find from the sample covariance matrix \(S=(s_{ij})\) the estimated factor loadings \({\hat \lambda}_{ij}\) of \(\lambda_{ij}\) and the estimated specific variance \({\hat \psi}_{ii}\) of \(\psi_{ii}\), \(i=1,...,p\); \(j=1,...,m\). The most theoretically appealing method for finding these estimates is that of maximum likelihood. So we discuss this method particularly.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4725536)