|
||||||||||
PREV CLASS NEXT CLASS | FRAMES NO FRAMES | |||||||||
SUMMARY: NESTED | FIELD | CONSTR | METHOD | DETAIL: FIELD | CONSTR | METHOD |
java.lang.Objectjml.manifold.Manifold
public class Manifold
Constructor Summary | |
---|---|
Manifold()
|
Method Summary | |
---|---|
static org.apache.commons.math.linear.RealMatrix |
adjacency(org.apache.commons.math.linear.RealMatrix X,
java.lang.String type,
double param,
java.lang.String distFunc)
Compute the symmetric adjacency matrix of the data set represented as a real data matrix X. |
static org.apache.commons.math.linear.RealMatrix |
adjacencyDirected(org.apache.commons.math.linear.RealMatrix X,
java.lang.String type,
double param,
java.lang.String distFunc)
Compute the directed adjacency matrix of the data set represented as a real data matrix X. |
static org.apache.commons.math.linear.RealMatrix |
calcLLR(org.apache.commons.math.linear.RealMatrix X,
double NN,
java.lang.String distFunc,
java.lang.String kernelType,
double kernelParam,
double lambda)
Compute local learning regularization matrix. |
static org.apache.commons.math.linear.RealMatrix |
cosine(org.apache.commons.math.linear.RealMatrix A,
org.apache.commons.math.linear.RealMatrix B)
Compute the cosine distance matrix between column vectors in matrix A and column vectors in matrix B. |
static org.apache.commons.math.linear.RealMatrix |
euclidean(org.apache.commons.math.linear.RealMatrix A,
org.apache.commons.math.linear.RealMatrix B)
Compute the Euclidean distance matrix between column vectors in matrix A and column vectors in matrix B. |
static org.apache.commons.math.linear.RealMatrix |
laplacian(org.apache.commons.math.linear.RealMatrix X,
java.lang.String type,
GraphOptions options)
Calculate the graph Laplacian of the adjacency graph of a data set represented as columns of a matrix X. |
static void |
main(java.lang.String[] args)
|
Methods inherited from class java.lang.Object |
---|
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
Constructor Detail |
---|
public Manifold()
Method Detail |
---|
public static void main(java.lang.String[] args)
args
- public static org.apache.commons.math.linear.RealMatrix laplacian(org.apache.commons.math.linear.RealMatrix X, java.lang.String type, GraphOptions options)
X
- data matrix with each column being a sampletype
- graph type, either "nn" or "epsballs"options
- data structure containing the following fields
NN - integer if type is "nn" (number of nearest neighbors),
or size of "epsballs"
DISTANCEFUNCTION - distance function used to make the graph
WEIGHTTYPPE = "binary" | "distance" | "heat" | "inner"
WEIGHTPARAM = width for heat kernel
NORMALIZE = 0 | 1 whether to return normalized graph Laplacian or not
public static org.apache.commons.math.linear.RealMatrix adjacency(org.apache.commons.math.linear.RealMatrix X, java.lang.String type, double param, java.lang.String distFunc)
X
- data matrix with each column being a feature vectortype
- graph type, either "nn" or "epsballs" ("eps")param
- integer if type is "nn", real number if type is "epsballs" ("eps")distFunc
- function mapping a (D x M) and a (D x N) matrix
to an M x N distance matrix (D: dimensionality)
either "euclidean" or "cosine"
public static org.apache.commons.math.linear.RealMatrix adjacencyDirected(org.apache.commons.math.linear.RealMatrix X, java.lang.String type, double param, java.lang.String distFunc)
X
- data matrix with each column being a feature vectortype
- graph type, either "nn" or "epsballs" ("eps")param
- integer if type is "nn", real number if type is "epsballs" ("eps")distFunc
- function mapping a (D x M) and a (D x N) matrix
to an M x N distance matrix (D: dimensionality)
either "euclidean" or "cosine"
public static org.apache.commons.math.linear.RealMatrix cosine(org.apache.commons.math.linear.RealMatrix A, org.apache.commons.math.linear.RealMatrix B)
A
- data matrix with each column being a feature vectorB
- data matrix with each column being a feature vector
public static org.apache.commons.math.linear.RealMatrix euclidean(org.apache.commons.math.linear.RealMatrix A, org.apache.commons.math.linear.RealMatrix B)
A
- data matrix with each column being a feature vectorB
- data matrix with each column being a feature vector
public static org.apache.commons.math.linear.RealMatrix calcLLR(org.apache.commons.math.linear.RealMatrix X, double NN, java.lang.String distFunc, java.lang.String kernelType, double kernelParam, double lambda)
X
- data matrix with each column being a feature vectorNN
- number of nearest neighbordistFunc
- function mapping a (D x M) and a (D x N) matrix
to an M x N distance matrix (D: dimensionality)
either "euclidean" or "cosine"kernelType
- 'linear' | 'poly' | 'rbf' | 'cosine'kernelParam
- -- | degree | sigma | --lambda
- graph regularization parameter
|
||||||||||
PREV CLASS NEXT CLASS | FRAMES NO FRAMES | |||||||||
SUMMARY: NESTED | FIELD | CONSTR | METHOD | DETAIL: FIELD | CONSTR | METHOD |