fmm-class: A class for creating a flat sparse model matrix...

In sakrejda/hierarchy: Package for handling inputs to hierarchical models.

Description

A class for creating a flat sparse model matrix representation

Fields

.n_row

number of observations in model matrix

.n_col

number of columns in the model matrix

.n_nze

number of non-zero entries in the model matrix

.nze

indexes of non-zero entries in the model matrix (RSC)

.start

where each row of the model matrix start in nze

.stop

where each row of the model matrix ends in nze

.xv

n_nze nze entries of the model matrix

.y

dependent data vector, if applicable

.groups

index into xv for each group of parameters produced by the formula

Methods

check_component(component)

Verify that the requested (formula) component is in the model matrix and return its name. If none is specified (NULL in calling function) then all are returned.

expose(...)

Extractor that takes a named vector and provides the relevant component with (optionally) a new name. The renaming syntax follows dplyr::rename so the new name is taken from the name of the argument and the element to extract is taken from the character vector content.

For example OBJ$expose(phi_n_col = 'n_col') would return the number of (implicit) model matrix columns with the name 'phi_n_col'. This is useful to construct lists that are going to be used in, e.g.-Stan.

get_data()

Get the data frame used to construct the matrix.

initialize(formula, data, configuration, N = nrow(data), ...)

Create the implicit mass matrix and store components.

list_components()

Get a list of the object's fields that can be exposed.

sakrejda/hierarchy documentation built on Aug. 23, 2019, 7:15 p.m.