class LayeredNetwork(NeuralNetwork):
# /**
# * Make a layered network
# */
def __init__(self):
self.hiddenLayers = []
self.inputLayer = Layer()
self.outputLayer = Layer()
self.links = [];
# /**
# * @see Network#getOutputValues()
# */
def getOutputValues(self):
#print "layeredNetwork.getOutputValues().self.outputLayer.getActivations().toString()" + str(self.outputLayer.getActivations().toString())
return self.outputLayer.getActivations()
# /**
# * @see Network#setInputValues(double[])
# */
def setInputValues(self, values):
if not isinstance(values, Vector):
raise TypeError('Need a Vector Type: ' + str(values))
self.inputLayer.setActivations(values)
# /**
# * Get the index of the node with the largest value
# * @return the index
# */
def getDiscreteOutputValue(self):
return self.outputLayer.getGreatestActivationIndex()
# /**
# * Get the binary output value
# * @return the binary output value
# */
def getBinaryOutputValue(self):
return self.outputLayer.getNode(0).getActivation() > .5
# /**
# * Get the input layer
# * @return the layer
# */
def getInputLayer(self):
return self.inputLayer
# /**
# * Get the list of middle layers
# * @return the list
# */
def getHiddenLayers(self):
return self.hiddenLayers
# /**
# * Get the output layer
# * @return the layer
# */
def getOutputLayer(self):
return self.outputLayer
# /**
# * Set the input layer
# * @param layer the layer
# */
def setInputLayer(self, layer):
self.inputLayer = layer
# /**
# * Set the output layer
# * @param layer the output layer
# */
def setOutputLayer(self, layer):
self.outputLayer = layer
# /**
# * Get the middle layer count
# * @return the middle layer count
# */
def getHiddenLayerCount(self):
return len(self.hiddenLayers)
# /**
# * Get the middle layer
# * @param i the index of the middle layer
# * @return the layer
# */
def getHiddenLayer(self, i):
#print "layeredNetwork.getHiddenLayer.self.hiddenLayers[i].__class__" + str(self.hiddenLayers[i].__dict__)
return self.hiddenLayers[i] # Returns a layer, dunno
# /**
# * Add a middle layer
# * @param layer the layer to add
# */
def addHiddenLayer(self, layer):
self.hiddenLayers.append(layer)
# /**
# * Disconnect this network
# */
def disconnect(self):
if (self.inputLayer != None and self.getHiddenLayerCount() > 0):
firstMiddle = self.getHiddenLayer(0)
self.inputLayer.disconnect(firstMiddle)
else:
if (self.inputLayer != None and self.outputLayer != None):
self.inputLayer.disconnect(self.outputLayer)
#for (i = 0; i + 1 < self.getHiddenLayerCount(); i++):
for i in range(self.getHiddenLayerCount() - 1):
first = self.getHiddenLayer(i)
second = self.getHiddenLayer(i + 1)
first.disconnect(second)
if (self.outputLayer != None and self.getHiddenLayerCount() > 0):
lastMiddle = getHiddenLayer(getHiddenLayerCount() - 1)
lastMiddle.disconnect(outputLayer)
# /**
# * Connect this network
# */
def connect(self):
if (self.inputLayer != None and self.getHiddenLayerCount() > 0):
firstMiddle = self.getHiddenLayer(0)
self.inputLayer.connect(firstMiddle)
else:
if (not self.inputLayer is None and not self.outputLayer is None):
self.inputLayer.connect(self.outputLayer)
#for (int i = 0; i < self.getHiddenLayerCount() - 1; i++): Did I implement right?
for i in range(self.getHiddenLayerCount() - 1):
first = self.getHiddenLayer(i)
second = self.getHiddenLayer(i + 1)
first.connect(second)
if (not self.outputLayer is None and self.getHiddenLayerCount() > 0):
lastMiddle = self.getHiddenLayer(self.getHiddenLayerCount() - 1)
lastMiddle.connect(self.outputLayer)
# /**
# * @see nn.NeuralNetwork#getLinks()
# */
def getLinks(self):
if not isinstance(self.links, list):
return self.links
self.links.append(self.inputLayer.getLinks())
for i in range(self.getHiddenLayerCount()):
self.links.append(self.getHiddenLayer(i).getLinks())
self.links.append(self.outputLayer.getLinks())
return self.links
def __init__(self):
self.hiddenLayers = []
self.inputLayer = Layer()
self.outputLayer = Layer()
self.links = [];