Wrapper function to generate t-SNE parameters

Wrapper function to generate t-SNE parameters

Usage

params_tsne(
  lr = NULL,
  n_epochs = 1000L,
  early_exag_iter = 250L,
  early_exag_factor = 12,
  late_exag_factor = NULL,
  theta = 0.5,
  n_interp_points = 3L,
  init = c("pca", "spectral", "random"),
  randomised = TRUE
)

Arguments

lr

Numeric or NULL. Learning rate. If NULL, the Rust backend sets it to max((n_samples / 12), 200), following the N-dependent heuristic of Belkina et al. (2019). Defaults to NULL.

n_epochs

Integer. Number of optimisation epochs. Defaults to 1000L.

early_exag_iter

Integer. Number of early exaggeration iterations. Defaults to 250L.

early_exag_factor

Numeric. Early exaggeration factor. Defaults to 12.0.

late_exag_factor

Numeric or NULL. Late exaggeration factor, if you wish to use one. Can be useful on large data sets (set it to 2.0 to 4.0). Defaults to NULL.

theta

Numeric. Barnes-Hut approximation angle. Lower values increase accuracy at the cost of speed. Defaults to 0.5.

n_interp_points

Integer. Number of interpolation points per grid cell for FFT acceleration. Defaults to 3L.

init

String. Embedding initialisation method. One of c("pca", "spectral", "random"). Defaults to "pca".

randomised

Boolean. Use randomised SVD for PCA initialisation. Defaults to TRUE.

Value

A named list with the following elements:

• lr - Numeric or NULL. Learning rate. If NULL, the Rust backend sets it to max((n_samples / 12), 200), following the N-dependent heuristic of Belkina et al. (2019). Defaults to NULL.

• n_epochs - Integer. Number of optimisation epochs. Defaults to 1000L.

• early_exag_iter - Integer. Number of early exaggeration iterations. Defaults to 250L.

• early_exag_factor - Numeric. Early exaggeration factor. Defaults to 12.0.

• late_exag_factor - Numeric or NULL. Late exaggeration factor, if you wish to use one. Can be useful on large data sets (set it to 2.0 to 4.0). Defaults to NULL.

• theta - Numeric. Barnes-Hut approximation angle. Lower values increase accuracy at the cost of speed. Defaults to 0.5.

• n_interp_points - Integer. Number of interpolation points per grid cell for FFT acceleration. Defaults to 3L.

• init - String. Embedding initialisation method. One of c("pca", "spectral", "random"). Defaults to "pca".

• randomised - Boolean. Use randomised SVD for PCA initialisation. Defaults to TRUE.

References

Belkina, et al., Nat. Commun., 2019