Estimating multivariate latent-structure models
Annals of Statistics
Institute of Mathematical Statistics
MetadataShow full item record
Bonhomme, S., Jochmans, K., & Robin, J. (2016). Estimating multivariate latent-structure models. Annals of Statistics, 44 (2), 540-563. https://doi.org/10.1214/15-AOS1376
© Institute of Mathematical Statistics, 2016. A constructive proof of identification of multilinear decompositions of multiway arrays is presented. It can be applied to show identification in a variety of multivariate latent structures. Examples are finite-mixture models and hidden Markov models. The key step to show identification is the joint diagonalization of a set of matrices in the same nonorthogonal basis. An estimator of the latent-structure model may then be based on a sample version of this joint-diagonalization problem. Algorithms are available for computation and we derive distribution theory. We further develop asymptotic theory for orthogonal-series estimators of component densities in mixture models and emission densities in hidden Markov models.
Supported by European Research Council Grant ERC-2010-StG-0263107-ENMUH. Supported by Sciences Po’s SAB grant “Nonparametric estimation of finite mixtures.” Supported by European Research Council Grant ERC-2010-AdG-269693-WASP and by Economic and Social Research Council Grant RES-589-28-0001 through the Centre for Microdata Methods and Practice.
External DOI: https://doi.org/10.1214/15-AOS1376
This record's URL: https://www.repository.cam.ac.uk/handle/1810/286881
Publisher's own licence