Skip to contents

Run UMAP Dimensionality Reduction

run_umap.Rd

A generic function to perform UMAP dimensionality reduction on feature data.

Usage

run_umap(x, ...)

# Default S3 method
run_umap(
  x,
  metadata_cols = NULL,
  scale = TRUE,
  n_neighbors = 15,
  n_components = 2,
  min_dist = 0.1,
  seed = 222,
  n_threads = NULL,
  verbose = TRUE,
  ...
)

# S3 method for class 'Sap'
run_umap(
  x,
  segment_type = c("motifs", "syllables", "bouts", "segments"),
  data_type = c("spectral_feature", "spectrogram", "traj_mat"),
  label = NULL,
  scale = TRUE,
  n_neighbors = 20,
  n_components = 2,
  min_dist = 0.1,
  seed = 222,
  n_threads = NULL,
  verbose = TRUE,
  ...
)

Arguments

x

An object to analyze, either a data frame or SAP object

...

Additional arguments passed to specific methods

metadata_cols

Column indices for metadata (for default method)

scale

Whether to scale features before UMAP

n_neighbors

Number of neighbors (default: 15)

n_components

Number of output dimensions (default: 2)

min_dist

Minimum distance parameter (default: 0.1)

seed

Random seed for reproducibility

n_threads

Number of computation threads

verbose

Whether to print progress messages

segment_type

For SAP objects: Type of segments to analyze ('motifs', 'syllables', 'bouts', 'segments')

data_type

For SAP objects: Type of feature data ('spectral_feature','spectrogram', 'traj_mat')

label

For SAP objects: Specific label to filter data

Value

For default method: Matrix of UMAP coordinates For SAP objects: Updated SAP object with UMAP coordinates stored in features slot

Details

This generic function supports UMAP analysis through two methods:

  • Default method for feature data frames

  • SAP object method for organized song features

Examples

if (FALSE) { # \dontrun{
# Run UMAP on feature data frame
coords <- run_umap(features,
                   metadata_cols = c(1:5),
                   n_neighbors = 15)

# Run UMAP on SAP object
sap_obj <- run_umap(sap_object,
                    segment_type = "motifs",
                    data_type = "spectral_feature")

# UMAP with specific parameters
coords <- run_umap(features,
                   metadata_cols = 1:3,
                   scale = TRUE,
                   n_neighbors = 20,
                   seed = 123)

# UMAP with label filtering
sap_obj <- run_umap(sap_obj,
                    segment_type = "syllables",
                    data_type = "spectral_feature",
                    label = "a")
} # }