Two-Stage Procrustes Rotation with Sparse Target Matrix and Least Squares Criterion with Regularization and Generalized Weighting

In factor analysis, a loading matrix is often rotated to simple target for its simplicity. For the purpose, Procrustes rotation minimizes discrepancy between and loadings using two types of approximation: 1) approximate zeros in by non-zeros loadings, 2) loadings. The central issue considered article that it equally treats approximation, while former more important simplifying matrix. Furthermore, well-known Simplimax computational inefficiency estimating sparse matrix, which yields considerable number local minima. research proposes new procedure consists following stages. first stage estimates with lesser cost regularization technique. second stage, target, emphasizing on approximation least squares criterion generalized weighing newly proposed study. simulation study real data examples revealed method surely simplifies matrices.