esaBcv_package {esaBcv}

esaBcv

Description

The esaBcv package provides functions to estimate the latent factors of a given matrix, no matter it is high-dimensional or not. It tries to first estimate the number of factors using Bi-cross-validation and then estimate the latent factor matrix and the noise variances using an Early-stopping-alternation method. The method is proposed by Art B. Owen and Jingshu Wang (2015).

Author(s)

Maintainer: Jingshu Wang <[e-mail]>

See Also

Owen and Wang (2015) Bi-cross-validation for factor analysis, http://arxiv.org/abs/1503.03515

Examples

## Not run: 
data(simdat)
result <- EsaBcv(simdat$Y)
plot(result)


## End(Not run)