Wrapper function for WNN parameters

Usage

params_sc_wnn(
  k_nn = 20L,
  knn_range = 200L,
  sigma_method = c("snn_farthest", "sigma_idx"),
  sigma_idx = 19L,
  snn_type = c("full_connection", "limited"),
  s_nn = 20L,
  sd_scale = 1,
  kernel_power = 1,
  cross_const = 1e-04,
  sigma_floor = 1e-08,
  knn = list()
)

Arguments

k_nn: Integer. Final number of multimodal neighbours per cell. Defaults to 20L.
knn_range: Integer. Candidate pool size per modality. Each cell's kNN input must contain at least this many neighbours. Defaults to 200L.
sigma_method: String. Bandwidth method. One of c("snn_farthest", "sigma_idx"). Defaults to "snn_farthest".
sigma_idx: Integer. "sigma_idx" only: 0-based kNN index for bandwidth.e. k_nn - 1). Defaults to 19L.
snn_type: String. sNN type. The limited version only considers edges that exist in the kNN. One of c("full_connection", "limited"). Defaults to "full_connection".
s_nn: Integer. "snn_farthest" only: kNN size used to build the SNN graph. Defaults to 20L.
sd_scale: Numeric. Multiplier on sigma. Defaults to 1.0.
kernel_power: Numeric. Kernel exponent power. Defaults to 1.0.
cross_const: Numeric. Cross-modality kernel stabiliser. Defaults to 1e-04.
sigma_floor: Numeric. Minimum sigma value (avoids division by zero). Defaults to 1e-08.
knn: List. Optional overrides for kNN parameters. See params_knn_defaults() for available parameters: k, knn_method, ann_dist, search_budget, n_trees, delta, diversify_prob, ef_budget, extract_knn, m, ef_construction, ef_search, n_list and n_probe. See params_knn_defaults() for the available elements. Defaults to list().

Value

A named list with the following elements:

k_nn - Integer. Final number of multimodal neighbours per cell. Defaults to 20L.
knn_range - Integer. Candidate pool size per modality. Each cell's kNN input must contain at least this many neighbours. Defaults to 200L.
sigma_method - String. Bandwidth method. One of c("snn_farthest", "sigma_idx"). Defaults to "snn_farthest".
sigma_idx - Integer. "sigma_idx" only: 0-based kNN index for bandwidth.e. k_nn - 1). Defaults to 19L.
snn_type - String. sNN type. The limited version only considers edges that exist in the kNN. One of c("full_connection", "limited"). Defaults to "full_connection".
s_nn - Integer. "snn_farthest" only: kNN size used to build the SNN graph. Defaults to 20L.
sd_scale - Numeric. Multiplier on sigma. Defaults to 1.0.
kernel_power - Numeric. Kernel exponent power. Defaults to 1.0.
cross_const - Numeric. Cross-modality kernel stabiliser. Defaults to 1e-04.
sigma_floor - Numeric. Minimum sigma value (avoids division by zero). Defaults to 1e-08.
The elements of params_knn_defaults(), overridden by knn, spliced in at this position.