
public class KernelMatrix extends Object
| Modifier and Type | Field and Description | 
|---|---|
(package private) Matrix | 
kernel
The kernel matrix 
 | 
| Constructor and Description | 
|---|
KernelMatrix(double[][] matrixArray)
Wraps the matrixArray in a KernelMatrix 
 | 
KernelMatrix(Matrix matrix)
Makes a new kernel matrix from matrix (with data copying). 
 | 
KernelMatrix(PrimitiveSimilarityFunction<? super O,DoubleDistance> kernelFunction,
            Relation<? extends O> database)
Deprecated. 
 
ID mapping is not reliable! 
 | 
KernelMatrix(PrimitiveSimilarityFunction<? super O,DoubleDistance> kernelFunction,
            Relation<? extends O> database,
            ArrayDBIDs ids)
Provides a new kernel matrix. 
 | 
| Modifier and Type | Method and Description | 
|---|---|
static Matrix | 
centerKernelMatrix(KernelMatrix kernelMatrix)
Centers the Kernel Matrix in Feature Space according to Smola et. 
 | 
static Matrix | 
centerMatrix(Matrix matrix)
Centers the matrix in feature space according to Smola et. 
 | 
double | 
getDistance(int o1,
           int o2)
Returns the kernel distance between the two specified objects. 
 | 
Matrix | 
getKernel()
Get the kernel matrix. 
 | 
double | 
getSimilarity(int o1,
             int o2)
Returns the kernel value of object o1 and object o2 
 | 
double | 
getSquaredDistance(int o1,
                  int o2)
Returns the squared kernel distance between the two specified objects. 
 | 
Matrix | 
getSubColumn(int i,
            List<Integer> ids)
Returns the ith kernel matrix column for all objects in ids 
 | 
Matrix | 
getSubMatrix(Collection<Integer> ids)
Returns a sub kernel matrix for all objects in ids 
 | 
String | 
toString()  | 
Matrix kernel
public KernelMatrix(double[][] matrixArray)
matrixArray - two dimensional double array@Deprecated public KernelMatrix(PrimitiveSimilarityFunction<? super O,DoubleDistance> kernelFunction, Relation<? extends O> database)
kernelFunction - the kernel function used to compute the kernel matrixdatabase - the database for which the kernel matrix is computedpublic KernelMatrix(PrimitiveSimilarityFunction<? super O,DoubleDistance> kernelFunction, Relation<? extends O> database, ArrayDBIDs ids)
kernelFunction - the kernel function used to compute the kernel matrixdatabase - the database that holds the objectsids - the IDs of those objects for which the kernel matrix is computedpublic KernelMatrix(Matrix matrix)
matrix - a matrixpublic double getDistance(int o1,
                 int o2)
o1 - first ObjectIDo2 - second ObjectIDpublic Matrix getKernel()
public double getSimilarity(int o1,
                   int o2)
o1 - ID of first objecto2 - ID of second objectpublic double getSquaredDistance(int o1,
                        int o2)
o1 - first ObjectIDo2 - second ObjectIDpublic Matrix getSubColumn(int i, List<Integer> ids)
i - the column which should be returnedids - the objectspublic Matrix getSubMatrix(Collection<Integer> ids)
ids - the objectspublic static Matrix centerMatrix(Matrix matrix)
centeredMatrix = centerKernelMatrix(uncenteredMatrix.copy()) {matrix - the matrix to be centeredpublic static Matrix centerKernelMatrix(KernelMatrix kernelMatrix)
centeredMatrix = centerKernelMatrix(uncenteredMatrix.copy()) {kernelMatrix - the kernel matrix to be centered