我正在通过《人工神经网络：实践课程》一书来学习神经网络，我正在构建一个神经网络来练习这本书，这本书对我说使用等于 -1 的偏差，我做到了，但我得到的结果对于某些输入是错误的，我不知道为什么，但是当我将偏差更改为 1 或其他数字正数时，我得到了所有输入的正确答案， 我认为这是因为是一个数字，将在每次与权重的交互中更新，这不应该发生，但正在发生，如果有人帮助我或尝试帮助我，我将不胜感激，对不起我的英语不好，这是我的代码：
#!/usr/bin/python3
从随机进口制服
input_train = [[-0.6508， 0.1097， 4.0009]， [-1.4492， 0.8896，4.4005]， [2.0850， 0.6876， 12.0710]， [0.2626， 1.1476， 7.7985]， [0.6418， 1.0234， 7.0427]， [0.2569, 0.6730, 8.3265], [1.1155, 0.6043, 7.4446], [0.0914, 0.3399, 7.0677], [0.0121, 0.5256, 4.6316], [-0.0429, 0.4660, 5.4323], [0.4340, 0.6870, 8.2287], [0.2735, 1.0287, 7.1934], [0.4839, 0.4851, 7.4850], [0.4089, -0.1267, 5.5019], [1.4391, 0.1614, 8.5843], [-0.9115, -0.1973, 2.1962], [0.3654, 1.0475, 7.4858], [0.2144, 0.7515, 7.1699], [0.2013, 1.0014, 6.5489], [0.6483, 0.2183, 5.8991], [-0.1147, 0.2242, 7.2435], [-0.7970, 0.8795, 3.8762], [-1.0625, 0.6366, 2.4707], [0.5307, 0.1285, 5.6883], [-1.2200, 0.7777, 1.7252], [0.3957， 0.1076， 5.6623]， [-0.1013， 0.5989， 7.1812]， [2.4482， 0.9455， 11.2095]， [2.0149， 0.6192， 10.9263]， [0.2012,0.2611,5.4631]]
output_train = [-1.0000， -1.0000， -1.0000， 1.0000，1.0000， -1.0000, 1.0000, -1.0000, 1.0000, 1.0000, -1.0000, 1.0000, -1.0000, -1.0000, -1.000, -1.0000, 1.0000, 1.0000, 1.0000, 1.0000, -1.0000, 1.0000, 1.0000, 1.0000, 1.0000, -1.0000
， -1.0000， 1.0000， -1.0000， 1.0000]类感知器： "感知器实现：Hebb规则">
def __init__(self, data, output, epoch, bias, learning_rate = 0.01):
self.data = data
self.output = output
self.epoch = epoch
self.bias = bias
self.learning_rate = learning_rate
self.n_samples = len(data)
self.n_atributtes = len(data[0])
self.weights = self.GenWeights()
def train(self):
for epoch in range(self.epoch):
erro = False
for row in range(self.n_samples):
result = self.Predict(self.data[row])
if result is not self.output[row]:
self.bias = self.UpdateBias(result, self.output[row])
for weights in range(self.n_atributtes):
self.weights[weights] = self.UpdateWeights(self.weights[weights], result, self.output[row], self.data[row][weights])
erro = True
if not erro: break
return
def Predict(self, datarow):
sum = self.bias
for ele in range(self.n_atributtes):
sum += datarow[ele] * self.weights[ele]
return 1.0000 if sum >= 0.0 else -1.0000
def UpdateWeights(self, weight, predict, expected, inputc):
return weight + (((expected - predict) * self.learning_rate) * inputc)
def UpdateBias(self, predict, expected):
return self.bias + (((expected - predict) * self.learning_rate) * -1 )
def GenWeights(self):
weights = list()
for atributtes in range(self.n_atributtes):
weights.insert(atributtes, uniform(0, 1))
return weights
def NewEntry(self, newrow):
sum = self.bias
for ele in range(self.n_atributtes):
sum += newrow[ele] * self.weights[ele]
return 1.0000 if sum >= 0.0 else -1.0000
def main((：
first_neural_network = perceptron(input_train, output_train, 1000, -1)
print("Initial Weights : {}" .format(first_neural_network.weights))
first_neural_network.train()
print("Trained Weights : {} and final bias : {} ".format(first_neural_network.weights, first_neural_network.bias))
a = 0
newlist = []
while True:
a = float(input("Enter With the first number : "))
newlist.append(a)
a = float(input("Enter with the second number : "))
newlist.append(a)
a = float(input("Enter with the third number : "))
newlist.append(a)
print(newlist)
print("result is : {0}", format(first_neural_network.NewEntry(newlist)))
newlist.clear()
del first_neural_network
return
如果名称== ">主要"： 主((
修改main函数以获取error rate
def main():
first_neural_network = perceptron(input_train, output_train, 1000, 1)
print("Initial Weights : {}" .format(first_neural_network.weights))
first_neural_network.train()
print("Trained Weights : {} and final bias : {} ".format(
first_neural_network.weights, first_neural_network.bias))
# a = 0
# newlist = []
# while True:
# a = float(input("Enter With the first number : "))
# newlist.append(a)
# a = float(input("Enter with the second number : "))
# newlist.append(a)
# a = float(input("Enter with the third number : "))
# newlist.append(a)
# print(newlist)
# print("result is : {0}".format(
# first_neural_network.NewEntry(newlist)))
# newlist.clear()
nb_error = 0
for newlist, expected in zip(input_train, output_train):
pred = first_neural_network.NewEntry(newlist)
print('true: {:2f} pred: {:2f}'.format(expected, pred))
nb_error += pred != expected
print('error rate%:{:.2f}'.format(float(nb_error)/first_neural_network.n_samples))
del first_neural_network
return
在您的配置中，error rate约为 0.47。 调整bias在指标中不发挥重要作用。 我假设您的数据不是线性分开的，您需要更复杂的模型来提高准确性