LLE2: Locally Linear Embedding

In kcf-jackson/maniTools: Manifold learning in R

Description

Locally Linear Embedding

Usage

LLE2(data, dim = 2, k)

Arguments

data	N x D matrix (N samples, D features)
dim	integer; The target dimension.
k	integer; Number of nearest neighbor.

Details

This is the exact same function as LLE in "RDRToolbox" except that the matrix inverse operation is replaced by the pseudoinverse operation since the former often has numerical issues. See "?RDRToolbox::LLE" for more details about usage.

Value

Projected data; N x dim matrix

References

Roweis, Sam T. and Saul, Lawrence K., "Nonlinear Dimensionality Reduction by Locally Linear Embedding", 2000

Examples

#Simulate data
sim_data <- swiss_roll(N = 600)
library(plotly)
p1 <- plotly_3D(sim_data); p1
lle_data <- LLE2(sim_data$data, dim = 2, k = 8)
p2 <- plotly_2D(lle_data, color = sim_data$colors); p2

kcf-jackson/maniTools documentation built on April 20, 2020, 8:29 a.m.