Build a relative NN-Descent index — rs_rnndescent_build • annsearchR

Use RnnDescentIndex instead.

Usage

rs_rnndescent_build(
  data,
  metric,
  s,
  r,
  t1,
  t2,
  n_trees,
  seed,
  precision,
  verbose
)

Arguments

data: Numeric matrix. Samples x features.
metric: String. One of c("euclidean", "cosine", "manhattan").
s: Integer. Neighbours sampled per node per local join.
r: Integer. Maximum out-degree after pruning.
t1: Integer. Outer iterations.
t2: Integer. Inner iterations per outer one.
n_trees: Integer or NULL. Random projection trees for the seed graph.
seed: Integer. Random seed.
precision: String. "float" or "double".
verbose: Boolean. Print progress.

Value

External pointer to the index.