from keras.datasets import boston_housing
(train_data, train_targets), (test_data, test_targets) = boston_housing.load_data()
# train_data.shape ##404个数据,13个特征
'''数据标准化'''
mean = train_data.mean(axis=0)
std = train_data.std(axis=0)
train_data = (train_data - mean) /std
test_data = (test_data - mean) /std
'''构建网络'''
from keras import models
from keras import layers
def build_model():
model = models.Sequential()
model.add(layers.Dense(64, activation='relu', input_shape=(train_data.shape[1],)))
model.add(layers.Dense(64, activation='relu'))
# 不能使用relu(relu会把负数变为0,正数保持不变)
# 标量回归(预测单一连续值的回归):最后一层不使用激活函数,可以预测任意范围内的值
model.add(layers.Dense(1))
# mse是均方误差,用于回归问题、交叉熵用于分类问题
# mae是平均绝对误差,表示预测值和目标值之差的绝对值,accuracy表示准确率(预测结果中准确的比重)
model.compile(optimizer='rmsprop', loss='mse', metrics=['mae'])
return model
'''k折交叉验证'''
# 因为数据集数据少,所以验证分数波动较大,因此用k折交叉验证以平均训练并取k个验证分数的平均值
import numpy as np
k = 4
num_val_samples = len(train_data) // k #整数除法
num_epochs = 100
all_mae_histories = []
## all_score_mae = []
for i in range(k):
print('processing fold #', i)
#依次把k分数据中的每一份作为校验数据集
val_data = train_data[i * num_val_samples : (i+1) * num_val_samples]
val_targets = train_targets[i* num_val_samples : (i+1) * num_val_samples]
#把剩下的k-1分数据作为训练数据,如果第i分数据作为校验数据,那么把前i-1份和第i份之后的数据连起来
partial_train_data = np.concatenate([train_data[: i * num_val_samples],
train_data[(i+1) * num_val_samples:]], axis = 0)
partial_train_targets = np.concatenate([train_targets[: i * num_val_samples],
train_targets[(i+1) * num_val_samples: ]],
axis = 0)
model = build_model()
'''
## model.fit(partial_train_data, partial_train_targets, epochs = num_epochs, batch_size = 1, verbose = 0)
# verbose表示静默训练
# 使用平均绝对误差代替准确率,平均绝对误差可以查看预测值和目标值之间差的绝对值
## val_mse_loss, val_mae = model.evaluate(val_data, val_targets, verbose = 0)
## all_score_mae.append(val_mae) # [2.56, 3.12, 3.18, 3.07]
## print(np.mean(all_score_mae))
'''
'''绘制MAE随epochs变化的图'''
# history获取验证集和训练集的accuracy/mae与损失值
history = model.fit(partial_train_data, partial_train_targets,
validation_data=(val_data, val_targets),
epochs = num_epochs,
batch_size = 1, verbose = 0)
mae_history = history.history['val_mean_absolute_error']
all_mae_histories.append(mae_history)
# 每轮一个数据,获取所有轮次中K折验证分数的MAE平均值
average_mae_history = [np.mean([x[i] for x in all_mae_histories]) for i in range(num_epochs)]
'''
绘制验证分数
每个数据的平均绝对误差MAE不应相差太大,应删去波动明显太大的数据
'''
import matplotlib.pyplot as plt
plt.plot(range(1, len(average_mae_history) + 1), average_mae_history)
plt.xlabel('Epochs')
plt.ylabel('Validation MAE')
plt.show()
'''删去前十个波动巨大的数据点
后面的数据段误差值变动很剧烈,但从图中很难看出不同点之间的差异究竟是多少
为了把第10个epoch后面的数据差异更明显的展现出来
对数据做一些变换(指数滑动平均),指数滑动平均具有把反复跳动的数据进行平滑的作用
def smooth_curve(points, factor=0.9):
smoothed_points = []
for point in points:
if smoothed_points:
previous = smoothed_points[-1]
smoothed_points.append(previous * factor + point * (1 - factor))
else:
smoothed_points.append(point)
return smoothed_points
smooth_mae_history = smooth_curve(average_mae_history[10:])
plt.plot(range(1, len(smooth_mae_history)+1), smooth_mae_history)
plt.xlabel('Epochs')
plt.ylabel('Validation MAE')
plt.show()
'''
# 训练最终模型
# epochs=30的时候,校验误差处于最低点,由此我们可以把循环训练的次数设定在30左右
# 经过校验数据的检测后,可以调整各项参数,然后再把模型重新训练一遍
# model = build_model()
# model.fit(train_data, train_targets, epochs = 30, batch_size = 16, verbose = 0)
# test_mse_score, test_mae_score = model.evaluate(test_data, test_targets)
# print(test_mae_score)