Confidence Areas for Fixed-Effects PCA

摘要: Principal component analysis (PCA) is often used to visualize data when the rows and columns are both of interest. In such a setting, there lack inferential methods on PCA output. We study asymptotic variance fixed-effects model for PCA, propose several approaches assessing variability estimates: method based parametric bootstrap, new cell-wise jackknife, as well computationally cheaper approximation jackknife. confidence regions by Procrustes rotation. Using simulation study, we compare proposed highlight strengths drawbacks each vary number rows, columns, strength relationships between variables.