public static class HiCO.Parameterizer<V extends NumberVector> extends AbstractParameterizer
Modifier and Type | Field and Description |
---|---|
static OptionID |
ALPHA_ID
The threshold for 'strong' eigenvectors: the 'strong' eigenvectors
explain a portion of at least alpha of the total variance.
|
(package private) double |
delta
Delta parameter
|
static OptionID |
DELTA_ID
Parameter to specify the threshold of a distance between a vector q and a
given space that indicates that q adds a new dimension to the space, must
be a double equal to or greater than 0.
|
private IndexFactory<V,FilteredLocalPCAIndex<NumberVector>> |
indexfactory
Factory to produce
|
static OptionID |
K_ID
Optional parameter to specify the number of nearest neighbors considered
in the PCA, must be an integer greater than 0.
|
(package private) int |
mu
Mu parameter
|
static OptionID |
MU_ID
Parameter to specify the smoothing factor, must be an integer greater
than 0.
|
Constructor and Description |
---|
Parameterizer() |
Modifier and Type | Method and Description |
---|---|
protected HiCO<V> |
makeInstance()
Make an instance after successful configuration.
|
protected void |
makeOptions(Parameterization config)
Add all options.
|
configure, make
public static final OptionID MU_ID
MU_ID
-nearest neighbor is used to compute the
correlation reachability of an object.
Key: -hico.mu
public static final OptionID K_ID
public static final OptionID DELTA_ID
Default value: 0.25
Key: -hico.delta
public static final OptionID ALPHA_ID
Default value: HiCO.DEFAULT_ALPHA
Key: -hico.alpha
int mu
double delta
private IndexFactory<V extends NumberVector,FilteredLocalPCAIndex<NumberVector>> indexfactory
protected void makeOptions(Parameterization config)
AbstractParameterizer
makeOptions
in class AbstractParameterizer
config
- Parameterization to add options to.protected HiCO<V> makeInstance()
AbstractParameterizer
makeInstance
in class AbstractParameterizer
Copyright © 2015 ELKI Development Team, Lehr- und Forschungseinheit für Datenbanksysteme, Ludwig-Maximilians-Universität München. License information.