Interface | Description |
---|---|
Disposable |
Provides a signature which enables wrapping classes to safely dispose
resources internally used by the implementing class - somewhat like
Closeable does for opened files.In GMLVQ implementation, it is assigned to classes which utilize ExecutorService so the internal thread pools can be shutdown safely. |
GMLVQCore.DefaultSettings |
Class | Description |
---|---|
GMLVQCore |
The implementation of the generalized matrix learning vector quantization.
|
GMLVQCore.Builder | |
GMLVQDefaultObserver |
Provides a light-weight, neutral, non-Weka implementation of the
Observer interface, so a Visualizer can be available even
when GMLVQCore was invoked from Tests or operated upon by
directly using its API. |
GradientDescent |
This class wraps the stochastic gradient descent of GMLVQ.
|
ProposedUpdate |
Each stochastic gradient descent composes a update which consists of updated
prototypes and an updated omega matrix (which defines how data points and
prototypes are mapped to the embedded space).
Most essential, this class provides the ProposedUpdate.incorporate(DataPoint) method which processes
individual data points selected by the GradientDescent and utilizes
their information to build the potential update. |
SigmoidFunction |
GMLVQ utilizes a sigmoid/Heaviside function to scale e.g.
|
UpdateManager |
The instance directing the learning process.
|