jml.subspace
Class LLE

java.lang.Object
  extended by jml.subspace.DimensionalityReduction
      extended by jml.subspace.LLE

public class LLE
extends DimensionalityReduction

Locally Linear Embedding (LLE).

Version:
1.0, Mar. 13th, 2013
Author:
Mingjie Qian

Field Summary
(package private)  int K
          Number of nearest neighbors to construct the neighborhood graph.
 
Fields inherited from class jml.subspace.DimensionalityReduction
r, R, X
 
Constructor Summary
LLE(int r)
          Constructor.
LLE(int r, int K)
          Constructor.
 
Method Summary
static void main(java.lang.String[] args)
           
 void run()
          Do dimensionality reduction.
static org.apache.commons.math.linear.RealMatrix run(org.apache.commons.math.linear.RealMatrix X, int K, int r)
          LLE (Locally Linear Embedding).
 
Methods inherited from class jml.subspace.DimensionalityReduction
feedData, feedData, getDataMatrix, getReducedDataMatrix, getReducedDimensionality, setReducedDimensionality
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

K

int K
Number of nearest neighbors to construct the neighborhood graph.

Constructor Detail

LLE

public LLE(int r)
Constructor.

Parameters:
r - number of dimensions to be reduced to

LLE

public LLE(int r,
           int K)
Constructor.

Parameters:
r - number of dimensions to be reduced to
K - number of nearest neighbors to construct the neighborhood graph
Method Detail

main

public static void main(java.lang.String[] args)
Parameters:
args -

run

public void run()
Description copied from class: DimensionalityReduction
Do dimensionality reduction.

Specified by:
run in class DimensionalityReduction

run

public static org.apache.commons.math.linear.RealMatrix run(org.apache.commons.math.linear.RealMatrix X,
                                                            int K,
                                                            int r)
LLE (Locally Linear Embedding).

Parameters:
X - a d x n data matrix
K - number of nearest neighbors
r - number of dimensions to be reduced to
Returns:
a r x n matrix which is the r dimensional representation of the given n examples