/**
* Neural network learns from Java-ML data set
*
* @param dataSetJML Dataset Java-ML data set
*/
@Override
public void buildClassifier(Dataset dataSetJML) {
DataSet dataSet =
JMLDataSetConverter.convertJMLToNeurophDataset(
dataSetJML, neuralNet.getInputsCount(), neuralNet.getOutputsCount());
neuralNet.learn(dataSet);
}
/**
* Prints network output for the each element from the specified training set.
*
* @param neuralNet neural network
* @param testSet test data set
*/
public static void testNeuralNetwork(NeuralNetwork neuralNet, DataSet testSet) {
for (DataSetRow testSetRow : testSet.getRows()) {
neuralNet.setInput(testSetRow.getInput());
neuralNet.calculate();
double[] networkOutput = neuralNet.getOutput();
System.out.print("Input: " + Arrays.toString(testSetRow.getInput()));
System.out.print(" Output: " + Arrays.toString(networkOutput));
}
}
public void orderForm(NeuralNetwork nnet) {
drawButton.setEnabled(false);
openInLauncherButton.setEnabled(false);
this.nnet = nnet;
graphTypeCombo.removeAllItems();
layerCombo.removeAllItems();
datasetTextField.setEditable(false);
nnetField.setEditable(false);
nnetField.setText(nnet.getLabel());
Layer[] layers = nnet.getLayers();
for (int i = layers.length; i > 0; i--) {
layerCombo.addItem(i);
}
}
private void fillAttributes() {
int layerIdx = nnet.getLayersCount() - 1; // output layerIdx by default
Attribute attr1;
Attribute attr2 = null;
if (layerCombo.isEnabled()) {
layerIdx = (Integer) layerCombo.getSelectedItem() - 1;
}
attr1 = new Attribute(layerIdx, false, "Layer");
if (attributeTxtField.isEnabled()) {
attribute = Integer.parseInt(attributeTxtField.getText());
attr2 = new Attribute(attribute, false, "Input");
}
graphBuilder.setAttribute1(attr1);
graphBuilder.setAttribute2(attr2);
outputNeuronCount = nnet.getLayerAt(layerIdx).getNeuronsCount();
}
@Override
public vafusion.recog.Character match(Object arg) {
BufferedImage charimg = (BufferedImage) arg;
OcrPlugin plugin = (OcrPlugin) nnet.getPlugin(OcrPlugin.OCR_PLUGIN_NAME);
return Character.mapCharacter(plugin.recognizeCharacter(charimg));
}
/** Sets default input and output neurons for network (first layer as input, last as output) */
public static void setDefaultIO(NeuralNetwork nnet) {
ArrayList<Neuron> inputNeuronsList = new ArrayList<Neuron>();
Layer firstLayer = nnet.getLayerAt(0);
for (Neuron neuron : firstLayer.getNeurons()) {
if (!(neuron instanceof BiasNeuron)) { // dont set input to bias
// neurons
inputNeuronsList.add(neuron);
}
}
Neuron[] inputNeurons = new Neuron[inputNeuronsList.size()];
inputNeurons = inputNeuronsList.toArray(inputNeurons);
Neuron[] outputNeurons = ((Layer) nnet.getLayerAt(nnet.getLayersCount() - 1)).getNeurons();
nnet.setInputNeurons(inputNeurons);
nnet.setOutputNeurons(outputNeurons);
}
/**
* Calculates predict values for every possible class that instance can be classified as that
*
* @param instnc Instance
* @return Map<Object, Double>
*/
@Override
public Map<Object, Double> classDistribution(Instance instnc) {
// Convert instance to double array
double[] item = convertInstanceToDoubleArray(instnc);
// set neural network input
neuralNet.setInput(item);
// calculate neural network output
neuralNet.calculate();
// find neuron with highest output
Map<Object, Double> possibilities = new HashMap<Object, Double>();
for (Neuron neuron : neuralNet.getOutputNeurons()) {
possibilities.put(neuron.getLabel(), neuron.getOutput());
}
return possibilities;
}
/**
* Classifies instance as one of possible classes
*
* @param instnc Instance to classify
* @return Object class as Object
*/
@Override
public Object classify(Instance instnc) {
double[] item = convertInstanceToDoubleArray(instnc);
// set neural network input
neuralNet.setInput(item);
// calculate neural network output
neuralNet.calculate();
// find neuron with highest output
Neuron maxNeuron = null;
double maxOut = Double.NEGATIVE_INFINITY;
for (Neuron neuron : neuralNet.getOutputNeurons()) {
if (neuron.getOutput() > maxOut) {
maxNeuron = neuron;
maxOut = neuron.getOutput();
}
}
// and return its label
return maxNeuron.getLabel();
}
private boolean validateAttribute(JTextField attributeField) {
if (attributeField.isEnabled()) {
try {
int a = Integer.parseInt(attributeField.getText().trim());
if (a > nnet.getInputsCount() || a <= 0) {
JOptionPane.showMessageDialog(
this,
"There is no data set attribute (input) with index: " + a,
"Error",
JOptionPane.ERROR_MESSAGE);
return false;
}
return true;
} catch (Exception e) {
JOptionPane.showMessageDialog(
this,
"Please enter an integer value in attribute field.",
"Error",
JOptionPane.ERROR_MESSAGE);
return false;
}
}
return true;
}
@Override
public void setNeuralNetwork(NeuralNetwork neuralNetwork) {
super.setNeuralNetwork(neuralNetwork);
int neuronsNum = neuralNetwork.getLayerAt(1).getNeuronsCount();
mapSize = (int) Math.sqrt(neuronsNum);
}
public ErrorSurface3D(NeuralNetwork nnet, DataSet dataset) {
dataProvider3D = new ErrorDataProvider3D(dataset, nnet);
dataSetRowCount = dataset.getRows().size();
outputNeuronCount = nnet.getOutputsCount();
}