flash_plot_structure: Create structure plot of factors or loadings for a flash fit
In willwerscheid/flashier: Empirical Bayes Matrix Factorization

flash_plot_structure

R Documentation

Create structure plot of factors or loadings for a flash fit

Description

Creates a “structure plot” (stacked bar plot) of posterior means for factors f_{jk} or loadings \ell_{ik}. Different “topics” or components (that is, the different factor/loadings pairs, as specified by kset) are represented by different colors. Values are normalized so that the maximum absolute value for each factor f_{\cdot k} or set of loadings \ell_{\cdot k} is equal to 1 and then stacked (see ldf.flash). Note that structure plots were designed for nonnegative loadings or “memberships”; if posterior means are not nonnegative then a different type of plot should be used (e.g., flash_plot_heatmap). By default, a 1-d embedding is used to arrange the rows i or columns j. This step is usually essential to creating a readable structure plot; for details, see structure_plot.

Usage

flash_plot_structure(
  fl,
  order_by_pve = FALSE,
  kset = NULL,
  pm_which = c("factors", "loadings"),
  pm_subset = NULL,
  pm_groups = NULL,
  pm_colors = NULL,
  gap = 1,
  ...
)

Arguments

`fl`	An object inheriting from class `flash`.
`order_by_pve`	If `order_by_pve = TRUE`, then factor/loadings pairs will be ordered according to proportion of variance explained, from highest to lowest. (By default, they are plotted in the same order as `kset`; or, if `kset` is `NULL`, then they are plotted in the same order as they are found in `fl`.)
`kset`	A vector of integers specifying the factor/loadings pairs to be plotted. If `order_by_pve = FALSE`, then `kset` also specifies the order in which they are to be plotted.
`pm_which`	Whether to plot loadings `L` or factors `F`.
`pm_subset`	A vector of row indices `i` or column indices `j` (depending on the argument to `pm_which`) specifying which values `\ell_{i \cdot}` or `f_{j \cdot}` are to be shown. If the dataset has row or column names, then names rather than indices may be specified. If `pm_subset = NULL`, then all values will be plotted.
`pm_groups`	A vector specifying the group to which each row of the data `y_{i \cdot}` or column `y_{\cdot j}` belongs (groups may be numeric indices or strings). A group must be provided for each plotted row `i` or column `j`, so that the length of `pm_groups` is exactly equal to the number of rows or columns in the full dataset or, if `pm_subset` is specified, in the subsetted dataset.
`pm_colors`	The colors of the “topics” or components (factor/loadings pairs).
`gap`	The horizontal spacing between groups. Ignored if `pm_groups` is not provided.
`...`	Additional parameters to be passed to `structure_plot`.