Data fusion using factor analysis and low-rank matrix completion

Data fusion involves the integration of multiple related datasets. The statistical file-matching problem is a canonical data in multivariate analysis, where objective to characterise joint distribution set variables when only strict subsets marginal distributions have been observed. Estimation covariance matrix full challenging given missing-data pattern. Factor analysis models use lower-dimensional latent data-generating process, and this introduces low-rank components complete-data population matrix. structure factor model can be exploited estimate from incomplete via completion. We prove identifiability under conditions on number factors shared over observed subsets. Additionally, we provide an EM algorithm for parameter estimation. On several real datasets, gives smaller reconstruction errors problems than common approaches