View Source Scholar.Neighbors.LargeVis (Scholar v0.3.0)

LargeVis algorithm for approximate k-nearest neighbor (k-NN) graph construction.

The algorithms works in the following way. First, the approximate k-NN graph is constructed using a random projection forest. Then, the graph is refined by looking at the neighbors of neighbors of every point for a fixed number of iterations. This step is called NN-expansion.

References

Visualizing Large-scale and High-dimensional Data.

Summary

Functions

expand(graph, tensor, opts)

fit(tensor, opts)

Constructs the approximate k-NN graph with LargeVis.

Functions

expand(graph, tensor, opts)

fit(tensor, opts)

Constructs the approximate k-NN graph with LargeVis.

Returns neighbor indices and distances.

Examples

iex> key = Nx.Random.key(12)
iex> tensor = Nx.iota({5, 2})
iex> {graph, distances} = Scholar.Neighbors.LargeVis.fit(tensor, num_neighbors: 2, metric: :squared_euclidean, min_leaf_size: 2, num_trees: 3, key: key)
iex> graph
#Nx.Tensor<
  u32[5][2]
  [
    [0, 1],
    [1, 0],
    [2, 1],
    [3, 2],
    [4, 3]
  ]
>
iex> distances
#Nx.Tensor<
  f32[5][2]
  [
    [0.0, 8.0],
    [0.0, 8.0],
    [0.0, 8.0],
    [0.0, 8.0],
    [0.0, 8.0]
  ]
>