def get_unit_values(self, features):
"""Calculate the activation of each unit in a neural network.
Args:
features: a vector of feature values
Returns:
units, a list of lists of numbers, where
units[a][b] is the activation of unit b in layer a
"""
units = list()
units.append(features)
for x in range(0, len(self.weights)):
layer = list()
for y in range(0, len(self.weights[x])):
v = common.dot(units[x], self.weights[x][y])
layer.append(self.activation(v))
units.append(layer)
# COMPLETE THIS IMPLEMENTATION
#print units
return units
def predict(self, features):
"""Return the prediction given perceptron weights on an example.
Args:
features: A vector of features, [f1, f2, ... fn], all numbers
Returns:
1 if w1 * f1 + w2 * f2 + ... * wn * fn + t > 0
0 otherwise
"""
if not len(self.weights) == len(features):
features = features + [1]
if common.dot(self.weights, features) > 0:
return 1
else:
return 0
def get_unit_values(self, features):
"""Calculate the activation of each unit in a neural network.
Args:
features: a vector of feature values
Returns:
units, a list of lists of numbers, where
units[a][b] is the activation of unit b in layer a
"""
# COMPLETE THIS IMPLEMENTATION
self.units = []
self.units.append (features)
for i in xrange (len (self.weights)):
self.units.append ([])
for layer in xrange (1, len (self.weights) + 1):
for unitWeights in self.weights[layer - 1]:
self.units[layer].append (self.activation (common.dot (unitWeights, self.units[layer - 1])))
return self.units
def predict(self, features):
"""Return the prediction given perceptron weights on an example.
Args:
features: A vector of features, [f1, f2, ... fn], all numbers
Returns:
1 if w1 * f1 + w2 * f2 + ... * wn * fn + t > 0
0 otherwise
"""
features.append(1)
ret = 0
value = common.dot(self.weights, features)
if(value >0):
ret = 1
del features[-1]
return ret
def predict(self, features):
"""Return the prediction given perceptron weights on an example.
Args:
features: A vector of features, [f1, f2, ... fn], all numbers
Returns:
1 if w1 * f1 + w2 * f2 + ... * wn * fn + t > 0
0 otherwise
"""
# COMPLETE THIS IMPLEMENTATION
# sum = 0
# for f, w in zip(features, self.weights):
# sum = sum + f * w
# features.append (1)
_tt = 0
if len(features) is not len(self.weights):
_tt = self.weights[len(self.weights) - 1]
self.weights.remove(_tt)
self.sum = common.dot(features, self.weights)
self.sum += _tt
return common.step(self.sum)
