Implments the traditional step-wise training of self-organized maps //TODO: finish this comment.
|
| TraditionalTraining (double initialWidth, double finalWidth, double initialRate, double finalRate, unsigned numIterations, NodeWeightInitialization *weightInitialization, NeighborhoodWindowFunction *windowFunc, Reporter *reporter) |
| Create a traditional SOM training algorithm that starts its learning rate and neighborhood width at initialWidth and initialRate then decreases them exponentially so that they both reach finalWidth and finalRate after numIterations iterations. Each iteration consists of one presentation of an input datum to the network and one weight update of the neighbors of the winning neuron at the current learning rate. More...
|
|
virtual | ~TraditionalTraining () |
|
GRand & | rand () |
| Return a reference to the pseudo-random number generator used by this object. More...
|
|
virtual void | train (GSelfOrganizingMap &map, const GMatrix *pIn) |
| Train the map. More...
|
|
virtual | ~TrainingAlgorithm () |
| Virtual destructor. More...
|
|
virtual GDomNode * | serialize (GDom *pDoc) const |
| Add this training algorithm to pDoc and return the resulting node Right now, default implementation is the only one there and it just adds an object with no fields. TODO: make serialize a pure virtual method and implement it in all the training algorithm subclasses. More...
|
|
Create a traditional SOM training algorithm that starts its learning rate and neighborhood width at initialWidth and initialRate then decreases them exponentially so that they both reach finalWidth and finalRate after numIterations iterations. Each iteration consists of one presentation of an input datum to the network and one weight update of the neighbors of the winning neuron at the current learning rate.
weightInitialization is the initialization function that will be used to initialize the node weights at the start of training. windowFunc is the window function used to determine the influence of neighbors on one another. reporter is the Reporter object that will be called to report progress during training.
The training object owns weightInialization, windowFunc, and reporter and so is responsible for deleting them.