super class for PLNPCAfamily and PLNnetworkfamily.
responsesthe matrix of responses common to every models
covariatesthe matrix of covariates common to every models
offsetsthe matrix of offsets common to every models
weightsthe vector of observation weights
inceptiona PLNfit object, obtained when no sparsifying penalty is applied.
modelsa list of PLNfit object, one per penalty.
criteriaa data frame with the values of some criteria (approximated log-likelihood, BIC, ICL, etc.) for the collection of models / fits BIC and ICL are defined so that they are on the same scale as the model log-likelihood, i.e. with the form, loglik - 0.5 penalty
convergencesends back a data frame with some convergence diagnostics associated with the optimization process (method, optimal value, etc)
new()Create a new PLNfamily object.
PLNfamily$new(responses, covariates, offsets, weights, control)getModel()Extract a model from a collection of models
varvalue of the parameter (rank for PLNPCA, sparsity for PLNnetwork) that identifies the model to be extracted from the collection. If no exact match is found, the model with closest parameter value is returned with a warning.
indexInteger index of the model to be returned. Only the first value is taken into account.
A PLNfit object
plot()Lineplot of selected criteria for all models in the collection
criteriaA valid model selection criteria for the collection of models. Includes loglik, BIC (all), ICL (PLNPCA) and pen_loglik, EBIC (PLNnetwork)
reverseA logical indicating whether to plot the value of the criteria in the "natural" direction (loglik - penalty) or in the "reverse" direction (-2 loglik + penalty). Default to FALSE, i.e use the natural direction, on the same scale as the log-likelihood.