Poisson Lognormal with Dimension Reduction for Count Matrices with Missing Data

---

Documentation for package ‘colvR’ version 0.2.1

• DESCRIPTION file.
• User guides, package vignettes and other documentation.
• Package NEWS.

Help Pages

BIC	BIC from the ELBO for PLN-PCA and ZIP-PLN-PCA models
covmat	Wrapper for 'lori::covmat'
Elbo_grad	ELBO and parameter gradients (ZIP‑PLN variational objective)
Elbo_grad_logS	Calculate the Elbo with log(S) instead of S and the gradients
IC	95% interval for a theta 95% confidence interval for theta
ICL	ICL from the ELBO for PLN-PCA and ZIP-PLN-PCA models
Init	Initialize parameters for a PLN-PCA with missing data
Init_ZIP	Initialize parameters for the ZIP case (zero-inflated)
Init_ZIP_q0	Init_ZIP_q0
Miss.PLNPCA	PLN-PCA with missing data (variational fit + imputation)
Miss.ZIPLNPCA	ZI‑PLN-PCA with missing data (variational fit + imputation)
Miss.ZIPLNPCA.logS	ZI‑PLN-PCA with missing data using log(S) parametrization
Miss.ZIPLNPCA_Steps	ZI‑PLN-PCA (missing data) with parameterwise optimization steps
Miss.ZIPLNPCA_VE	ZI‑PLN-PCA (missing data) — variational E‑step solver
Predictions	Predictive simulations and intervals
select_model_bic	Model selection over candidate latent dimensions (BIC or ICL)
Simul	Simulate zero-inflated Poisson log-normal PCA data
V_theta	Variance-covariance matrix of theta
waterbird	Waterbird count data