Plots features weight distribution — plot_features_weight

Plots the features weight or importance per dataset and latent dimension.

Usage

plot_features_weight_distr(
  method_output,
  latent_dimensions = NULL,
  datasets = NULL,
  features_metric = c("signed_importance", "weight", "importance"),
  top_n = 0,
  mo_data = NULL,
  label_cols = NULL,
  truncate = NULL,
  text_size = 2.5
)

Arguments

method_output: Integration method output generated via the get_output() function.
latent_dimensions: Character vector giving the latent dimensions to display. Default value is NULL, i.e. all latent dimensions will be shown.
datasets: Character vector giving the datasets to display. Default value is NULL, i.e. all datasets will be shown.
features_metric: Character, which attribute should be plotted: can be 'signed_importance' (i.e. importance value but with the weight sign), 'importance' or 'weight'. Default value is 'signed_importance'.
top_n: Integer, number of top features (in terms of importance) for which the label should be shown. Default value is 0.
mo_data: A MultiDataSet object. Only used if label_cols is not NULL.
label_cols: Character or named list of character, giving for each dataset the name of the column in the corresponding features metadata to use as label. If one value, will be used for all datasets. If list, the names must correspond to the names of the datasets in mo_data. If a dataset is missing from the list or no value is provided, feature IDs will be used as labels. Alternatively, use feature_id to get the feature IDs as labels.
truncate: Integer, width to which the labels should be truncated (to avoid issues with very long labels in plots). If NULL (default value), no truncation will be performed.
text_size: Numeric, size of the feature labels.

Value

A patchwork of plots.