bayespca: bayespca: Regularized Principal Component Analysis via...

In davidevdt/bayespca: A Package for Estimating PCA with Variational Bayes Inference

Description

A package for estimating PCA with Variational Bayes inference.

Details

Bayesian estimation of weight vectors in PCA. To achieve regularization, the method allows specifying fixed variances in the prior distributions of the weights; alternatively, it is possible to implement Jeffrey's and Inverse Gamma priors on such parameters. In turn, the Inverse Gamma's can have fixed shape hyperparameter; and fixed or random scale hyperparameter. Last, the method allows performing component-specific stochastic variable selection ('spike-and-slab' prior).

Functions

• vbpca for model estimation;

• vbpca_control for settings of control parameters;

• is.vbpca for testing the class;

• plothpdi for plotting high probability density intervals.

Author(s)

D. Vidotto <[e-mail]>

References

• [1] C. M. Bishop. 'Variational PCA'. In Proc. Ninth Int. Conf. on Artificial Neural Networks. ICANN, 1999.

• [2] E. I. George, R. E. McCulloch (1993). 'Variable Selection via Gibbs Sampling'. Journal of the American Statistical Association (88), 881-889.

davidevdt/bayespca documentation built on Dec. 5, 2020, 3:28 a.m.