GClasses
|
Isomap is a manifold learning algorithm that uses the Floyd-Warshall algorithm to compute an estimate of the geodesic distance between every pair of points using local neighborhoods, and then uses classic multidimensional scaling to compute a low-dimensional projection. More...
#include <GManifold.h>
Public Member Functions | |
GIsomap (size_t neighborCount, size_t targetDims, GRand *pRand) | |
GIsomap (GDomNode *pNode, GLearnerLoader &ll) | |
virtual | ~GIsomap () |
GDomNode * | serialize (GDom *pDoc) |
Serializes this object. | |
void | dropDisconnectedPoints () |
If there are any points that are not connected to the main group, just drop them instead of throwing an exception. (Note that this may cause the results to contain a different number of rows than the input.) | |
void | setNeighborFinder (GNeighborFinder *pNF) |
Specifies to use the neighborhoods determined by the specified neighbor-finder instead of the nearest Euclidean-distance neighbors to establish local linearity. If this method is called, it will also use the number of neighbors and the data associated with pNF, and ignore the number of neighbors specified to the constructor, and ignore the data passed to the "transform" method. | |
virtual GMatrix * | doit (GMatrix &in) |
Performs NLDR. | |
Protected Attributes | |
size_t | m_neighborCount |
size_t | m_targetDims |
GNeighborFinder * | m_pNF |
GRand * | m_pRand |
bool | m_dropDisconnectedPoints |
Isomap is a manifold learning algorithm that uses the Floyd-Warshall algorithm to compute an estimate of the geodesic distance between every pair of points using local neighborhoods, and then uses classic multidimensional scaling to compute a low-dimensional projection.
GClasses::GIsomap::GIsomap | ( | size_t | neighborCount, |
size_t | targetDims, | ||
GRand * | pRand | ||
) |
GClasses::GIsomap::GIsomap | ( | GDomNode * | pNode, |
GLearnerLoader & | ll | ||
) |
virtual GClasses::GIsomap::~GIsomap | ( | ) | [virtual] |
Performs NLDR.
Implements GClasses::GTransform.
void GClasses::GIsomap::dropDisconnectedPoints | ( | ) | [inline] |
If there are any points that are not connected to the main group, just drop them instead of throwing an exception. (Note that this may cause the results to contain a different number of rows than the input.)
void GClasses::GIsomap::setNeighborFinder | ( | GNeighborFinder * | pNF | ) |
Specifies to use the neighborhoods determined by the specified neighbor-finder instead of the nearest Euclidean-distance neighbors to establish local linearity. If this method is called, it will also use the number of neighbors and the data associated with pNF, and ignore the number of neighbors specified to the constructor, and ignore the data passed to the "transform" method.
bool GClasses::GIsomap::m_dropDisconnectedPoints [protected] |
size_t GClasses::GIsomap::m_neighborCount [protected] |
GNeighborFinder* GClasses::GIsomap::m_pNF [protected] |
GRand* GClasses::GIsomap::m_pRand [protected] |
size_t GClasses::GIsomap::m_targetDims [protected] |