【深度学习】实验13 使用Dropout抑制过拟合

文章目录

  • 使用Dropout抑制过拟合
    • 1. 环境准备
    • 2. 导入数据集
    • 3. 对所有数据的预测
      • 3.1 数据集
      • 3.2 构建神经网络
    • 3.3 训练模型
      • 3.4 分析模型
    • 4. 对未见过数据的预测
      • 4.1 划分数据集
      • 4.2 构建神经网络
      • 4.3 训练模型
      • 4.4 分析模型
    • 5. 使用Dropout抑制过拟合
      • 5.1 构建神经网络
      • 5.2 训练模型
      • 5.3 分析模型
    • 6. 正则化
      • 6.1 神经网络太过复杂容易过拟合
      • 6.2 太简单容易欠拟合
      • 6.3 选取适当的神经网络
  • 附:系列文章

使用Dropout抑制过拟合

Dropout是一种常用的神经网络正则化方法,主要用于防止过拟合。在深度学习中,由于网络层数过多,参数数量庞大,模型容易过拟合,并且在测试时产生较大的泛化误差。Dropout方法借鉴了集成学习中的Bagging思想,通过随机的方式,将一部分神经元的输出设置为0,从而减少过拟合的可能。

Dropout方法最早由Hinton等人提出,其基本思想是在训练时,以一定的概率随机地将网络中某些神经元的输出置为0。这种随机的行为可以被看作是一种对网络进行了部分剪枝,从而增加了网络的容忍性,使网络更加健壮,同时也避免了网络中某些特定的神经元对整个网络的过度依赖。

Dropout方法的具体实现如下:在每次训练过程中,以一定的概率p随机选择一部分神经元并将其置为0,被选择的神经元不参与后续的训练和反向传播。在测试时,为了保持模型的稳定性和一致性,一般不会采取随机化的方式,而是将每个神经元的权重乘以概率p,这里的p是在训练时选择的那个概率。

Dropout方法不仅可用于全连接网络,也可用于卷积神经网络和循环神经网络中,以减少过拟合现象。并且,它的实现简单,仅需要在模型训练时对每个神经元以概率p随机地进行挑选和保留,所以Dropout方法得到了广泛的应用和推广。

总之,Dropout方法可以在一定程度上提高模型的准确性和泛化能力,对于防止过拟合有着较好的效果。但是需要注意的是,Dropout方法会导致训练过程中每个mini-batch的梯度都不同,所以在使用Dropout方法时需要调整学习率,以保证模型的收敛速度和效果。

1. 环境准备

# 导入库
import keras
from keras import layers
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
Using TensorFlow backend.

2. 导入数据集

# 导入数据集
data = pd.read_csv('./dataset/credit-a.csv', header=None)
data
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0 0 30.83 0.000 0 0 9 0 1.250 0 0 1 1 0 202 0.0 -1
1 1 58.67 4.460 0 0 8 1 3.040 0 0 6 1 0 43 560.0 -1
2 1 24.50 0.500 0 0 8 1 1.500 0 1 0 1 0 280 824.0 -1
3 0 27.83 1.540 0 0 9 0 3.750 0 0 5 0 0 100 3.0 -1
4 0 20.17 5.625 0 0 9 0 1.710 0 1 0 1 2 120 0.0 -1
5 0 32.08 4.000 0 0 6 0 2.500 0 1 0 0 0 360 0.0 -1
6 0 33.17 1.040 0 0 7 1 6.500 0 1 0 0 0 164 31285.0 -1
7 1 22.92 11.585 0 0 2 0 0.040 0 1 0 1 0 80 1349.0 -1
8 0 54.42 0.500 1 1 5 1 3.960 0 1 0 1 0 180 314.0 -1
9 0 42.50 4.915 1 1 9 0 3.165 0 1 0 0 0 52 1442.0 -1
10 0 22.08 0.830 0 0 0 1 2.165 1 1 0 0 0 128 0.0 -1
11 0 29.92 1.835 0 0 0 1 4.335 0 1 0 1 0 260 200.0 -1
12 1 38.25 6.000 0 0 5 0 1.000 0 1 0 0 0 0 0.0 -1
13 0 48.08 6.040 0 0 5 0 0.040 1 1 0 1 0 0 2690.0 -1
14 1 45.83 10.500 0 0 8 0 5.000 0 0 7 0 0 0 0.0 -1
15 0 36.67 4.415 1 1 5 0 0.250 0 0 10 0 0 320 0.0 -1
16 0 28.25 0.875 0 0 6 0 0.960 0 0 3 0 0 396 0.0 -1
17 1 23.25 5.875 0 0 8 0 3.170 0 0 10 1 0 120 245.0 -1
18 0 21.83 0.250 0 0 1 1 0.665 0 1 0 0 0 0 0.0 -1
19 1 19.17 8.585 0 0 2 1 0.750 0 0 7 1 0 96 0.0 -1
20 0 25.00 11.250 0 0 0 0 2.500 0 0 17 1 0 200 1208.0 -1
21 0 23.25 1.000 0 0 0 0 0.835 0 1 0 1 2 300 0.0 -1
22 1 47.75 8.000 0 0 0 0 7.875 0 0 6 0 0 0 1260.0 -1
23 1 27.42 14.500 0 0 10 1 3.085 0 0 1 1 0 120 11.0 -1
24 1 41.17 6.500 0 0 8 0 0.500 0 0 3 0 0 145 0.0 -1
25 1 15.83 0.585 0 0 0 1 1.500 0 0 2 1 0 100 0.0 -1
26 1 47.00 13.000 0 0 3 2 5.165 0 0 9 0 0 0 0.0 -1
27 0 56.58 18.500 0 0 1 2 15.000 0 0 17 0 0 0 0.0 -1
28 0 57.42 8.500 0 0 11 1 7.000 0 0 3 1 0 0 0.0 -1
29 0 42.08 1.040 0 0 9 0 5.000 0 0 6 0 0 500 10000.0 -1
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
623 1 28.58 3.750 0 0 0 0 0.250 1 0 1 0 0 40 154.0 1
624 0 22.25 9.000 0 0 12 0 0.085 1 1 0 1 0 0 0.0 1
625 0 29.83 3.500 0 0 0 0 0.165 1 1 0 1 0 216 0.0 1
626 1 23.50 1.500 0 0 9 0 0.875 1 1 0 0 0 160 0.0 1
627 0 32.08 4.000 1 1 2 0 1.500 1 1 0 0 0 120 0.0 1
628 0 31.08 1.500 1 1 9 0 0.040 1 1 0 1 2 160 0.0 1
629 0 31.83 0.040 1 1 6 0 0.040 1 1 0 1 0 0 0.0 1
630 1 21.75 11.750 0 0 0 0 0.250 1 1 0 0 0 180 0.0 1
631 1 17.92 0.540 0 0 0 0 1.750 1 0 1 0 0 80 5.0 1
632 0 30.33 0.500 0 0 1 1 0.085 1 1 0 0 2 252 0.0 1
633 0 51.83 2.040 1 1 13 7 1.500 1 1 0 1 0 120 1.0 1
634 0 47.17 5.835 0 0 9 0 5.500 1 1 0 1 0 465 150.0 1
635 0 25.83 12.835 0 0 2 0 0.500 1 1 0 1 0 0 2.0 1
636 1 50.25 0.835 0 0 12 0 0.500 1 1 0 0 0 240 117.0 1
637 1 37.33 2.500 0 0 3 1 0.210 1 1 0 1 0 260 246.0 1
638 1 41.58 1.040 0 0 12 0 0.665 1 1 0 1 0 240 237.0 1
639 1 30.58 10.665 0 0 8 1 0.085 1 0 12 0 0 129 3.0 1
640 0 19.42 7.250 0 0 6 0 0.040 1 0 1 1 0 100 1.0 1
641 1 17.92 10.210 0 0 13 7 0.000 1 1 0 1 0 0 50.0 1
642 1 20.08 1.250 0 0 0 0 0.000 1 1 0 1 0 0 0.0 1
643 0 19.50 0.290 0 0 5 0 0.290 1 1 0 1 0 280 364.0 1
644 0 27.83 1.000 1 1 1 1 3.000 1 1 0 1 0 176 537.0 1
645 0 17.08 3.290 0 0 3 0 0.335 1 1 0 0 0 140 2.0 1
646 0 36.42 0.750 1 1 1 0 0.585 1 1 0 1 0 240 3.0 1
647 0 40.58 3.290 0 0 6 0 3.500 1 1 0 0 2 400 0.0 1
648 0 21.08 10.085 1 1 11 1 1.250 1 1 0 1 0 260 0.0 1
649 1 22.67 0.750 0 0 0 0 2.000 1 0 2 0 0 200 394.0 1
650 1 25.25 13.500 1 1 13 7 2.000 1 0 1 0 0 200 1.0 1
651 0 17.92 0.205 0 0 12 0 0.040 1 1 0 1 0 280 750.0 1
652 0 35.00 3.375 0 0 0 1 8.290 1 1 0 0 0 0 0.0 1

653 rows × 16 columns

data.iloc[:, -1].unique()
array([-1,  1])

3. 对所有数据的预测

3.1 数据集

x = data.iloc[:, :-1].values
y = data.iloc[:, -1].replace(-1, 0).values.reshape(-1, 1)
x.shape, y.shape
((653, 15), (653, 1))

3.2 构建神经网络

model = keras.Sequential()
model.add(layers.Dense(128, input_dim=15, activation='relu'))
model.add(layers.Dense(128, activation='relu'))
model.add(layers.Dense(128, activation='relu'))
model.add(layers.Dense(1, activation='sigmoid'))
model.summary()
Model: "sequential_1"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
dense_1 (Dense)              (None, 128)               2048      
_________________________________________________________________
dense_2 (Dense)              (None, 128)               16512     
_________________________________________________________________
dense_3 (Dense)              (None, 128)               16512     
_________________________________________________________________
dense_4 (Dense)              (None, 1)                 129       
=================================================================
Total params: 35,201
Trainable params: 35,201
Non-trainable params: 0
_________________________________________________________________
model.compile(optimizer='adam',
              loss='binary_crossentropy',
              metrics=['acc']
)
WARNING:tensorflow:From /home/nlp/anaconda3/lib/python3.7/site-packages/tensorflow/python/ops/nn_impl.py:180: add_dispatch_support..wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.where in 2.0, which has the same broadcast rule as np.where

3.3 训练模型

history = model.fit(x, y, epochs=1000)
WARNING:tensorflow:From /home/nlp/anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:422: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.

Epoch 1/1000
653/653 [==============================] - 0s 434us/step - loss: 7.5273 - acc: 0.5988
Epoch 2/1000
653/653 [==============================] - 0s 92us/step - loss: 3.7401 - acc: 0.6187
Epoch 3/1000
653/653 [==============================] - 0s 75us/step - loss: 3.6464 - acc: 0.5712
Epoch 4/1000
653/653 [==============================] - 0s 56us/step - loss: 10.2291 - acc: 0.6631
Epoch 5/1000
653/653 [==============================] - 0s 63us/step - loss: 2.0400 - acc: 0.6233
Epoch 6/1000
653/653 [==============================] - 0s 120us/step - loss: 2.4279 - acc: 0.6217
Epoch 7/1000
653/653 [==============================] - 0s 105us/step - loss: 2.3289 - acc: 0.6325
Epoch 8/1000
653/653 [==============================] - 0s 159us/step - loss: 3.2521 - acc: 0.6294
Epoch 9/1000
653/653 [==============================] - 0s 89us/step - loss: 2.6005 - acc: 0.6294
Epoch 10/1000
653/653 [==============================] - 0s 118us/step - loss: 1.3997 - acc: 0.6738
……
Epoch 1000/1000
653/653 [==============================] - 0s 106us/step - loss: 0.2630 - acc: 0.9326

3.4 分析模型

history.history.keys()
dict_keys(['loss', 'acc'])
plt.plot(history.epoch, history.history.get('loss'), c='r')
plt.plot(history.epoch, history.history.get('acc'), c='b')
[]

【深度学习】实验13 使用Dropout抑制过拟合_第1张图片

4. 对未见过数据的预测

4.1 划分数据集

x_train = x[:int(len(x)*0.75)]
x_test = x[int(len(x)*0.75):]
y_train = y[:int(len(x)*0.75)]
y_test = y[int(len(x)*0.75):]
x_train.shape, x_test.shape, y_train.shape, y_test.shape
((489, 15), (164, 15), (489, 1), (164, 1))

4.2 构建神经网络

model = keras.Sequential()
model.add(layers.Dense(128, input_dim=15, activation='relu'))
model.add(layers.Dense(128, activation='relu'))
model.add(layers.Dense(128, activation='relu'))
model.add(layers.Dense(1, activation='sigmoid'))
#admam:利用梯度的一阶矩估计和二阶矩估计动态调整每个参数的学习率.
model.compile(optimizer='adam',
              loss='binary_crossentropy',
              metrics=['acc']
)

4.3 训练模型

history = model.fit(x_train, y_train, epochs=1000, validation_data=(x_test, y_test))
Train on 489 samples, validate on 164 samples
Epoch 1/1000
489/489 [==============================] - 0s 487us/step - loss: 14.4564 - acc: 0.5951 - val_loss: 3.3778 - val_acc: 0.7256
Epoch 2/1000
489/489 [==============================] - 0s 110us/step - loss: 6.0909 - acc: 0.6012 - val_loss: 2.0924 - val_acc: 0.7195
Epoch 3/1000
489/489 [==============================] - 0s 195us/step - loss: 2.4527 - acc: 0.6074 - val_loss: 1.0763 - val_acc: 0.7378
Epoch 4/1000
489/489 [==============================] - 0s 183us/step - loss: 1.0751 - acc: 0.6585 - val_loss: 0.8990 - val_acc: 0.7134
Epoch 5/1000
489/489 [==============================] - 0s 155us/step - loss: 1.2669 - acc: 0.6503 - val_loss: 1.8094 - val_acc: 0.6585
Epoch 6/1000
489/489 [==============================] - 0s 202us/step - loss: 3.6742 - acc: 0.6892 - val_loss: 1.1836 - val_acc: 0.4573
Epoch 7/1000
489/489 [==============================] - 0s 166us/step - loss: 1.7544 - acc: 0.7301 - val_loss: 2.0060 - val_acc: 0.4573
Epoch 8/1000
489/489 [==============================] - 0s 185us/step - loss: 1.4768 - acc: 0.6605 - val_loss: 0.8917 - val_acc: 0.5427
Epoch 9/1000
489/489 [==============================] - 0s 163us/step - loss: 1.6829 - acc: 0.6667 - val_loss: 4.7695 - val_acc: 0.4573
Epoch 10/1000
489/489 [==============================] - 0s 157us/step - loss: 8.4323 - acc: 0.7239 - val_loss: 2.0879 - val_acc: 0.7439
……
Epoch 1000/1000
489/489 [==============================] - 0s 97us/step - loss: 0.0272 - acc: 0.9877 - val_loss: 2.2746 - val_acc: 0.8049

4.4 分析模型

history.history.keys()
dict_keys(['val_loss', 'val_acc', 'loss', 'acc'])
plt.plot(history.epoch, history.history.get('val_acc'), c='r', label='val_acc')
plt.plot(history.epoch, history.history.get('acc'), c='b', label='acc')
plt.legend()

【深度学习】实验13 使用Dropout抑制过拟合_第2张图片

model.evaluate(x_train, y_train)
489/489 [==============================] - 0s 37us/step
[0.021263938083038197, 0.9897750616073608]
model.evaluate(x_test, y_test)
164/164 [==============================] - 0s 46us/step
[2.274633582976715, 0.8048780560493469]

过拟合:在训练数据正确率非常高, 在测试数据上比较低

5. 使用Dropout抑制过拟合

5.1 构建神经网络

model = keras.Sequential()
model.add(layers.Dense(128, input_dim=15, activation='relu'))
model.add(layers.Dropout(0.5))
model.add(layers.Dense(128, activation='relu'))
model.add(layers.Dropout(0.5))
model.add(layers.Dense(128, activation='relu'))
model.add(layers.Dropout(0.5))
model.add(layers.Dense(1, activation='sigmoid'))
model.summary()
Model: "sequential_3"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
dense_9 (Dense)              (None, 128)               2048      
_________________________________________________________________
dropout_1 (Dropout)          (None, 128)               0         
_________________________________________________________________
dense_10 (Dense)             (None, 128)               16512     
_________________________________________________________________
dropout_2 (Dropout)          (None, 128)               0         
_________________________________________________________________
dense_11 (Dense)             (None, 128)               16512     
_________________________________________________________________
dropout_3 (Dropout)          (None, 128)               0         
_________________________________________________________________
dense_12 (Dense)             (None, 1)                 129       
=================================================================
Total params: 35,201
Trainable params: 35,201
Non-trainable params: 0
________________________________________________________________
model.compile(optimizer='adam',
              loss='binary_crossentropy',
              metrics=['acc']
)

5.2 训练模型

history = model.fit(x_train, y_train, epochs=1000, validation_data=(x_test, y_test))
Train on 489 samples, validate on 164 samples
Epoch 1/1000
489/489 [==============================] - 1s 1ms/step - loss: 41.6885 - acc: 0.5378 - val_loss: 9.9666 - val_acc: 0.6768
Epoch 2/1000
489/489 [==============================] - 0s 298us/step - loss: 53.1358 - acc: 0.5358 - val_loss: 11.0265 - val_acc: 0.6951
Epoch 3/1000
489/489 [==============================] - 0s 173us/step - loss: 36.9899 - acc: 0.5828 - val_loss: 11.6578 - val_acc: 0.6890
Epoch 4/1000
489/489 [==============================] - 0s 177us/step - loss: 43.3404 - acc: 0.5808 - val_loss: 7.5652 - val_acc: 0.6890
Epoch 5/1000
489/489 [==============================] - 0s 197us/step - loss: 23.3085 - acc: 0.6196 - val_loss: 7.9913 - val_acc: 0.6890
Epoch 6/1000
489/489 [==============================] - 0s 254us/step - loss: 24.1833 - acc: 0.6155 - val_loss: 5.5747 - val_acc: 0.7073
Epoch 7/1000
489/489 [==============================] - 0s 229us/step - loss: 19.7051 - acc: 0.5890 - val_loss: 5.5711 - val_acc: 0.7012
Epoch 8/1000
489/489 [==============================] - 0s 180us/step - loss: 22.1131 - acc: 0.5849 - val_loss: 7.0290 - val_acc: 0.6890
Epoch 9/1000
489/489 [==============================] - 0s 172us/step - loss: 23.2305 - acc: 0.6115 - val_loss: 4.2624 - val_acc: 0.6951
Epoch 10/1000
……
Epoch 1000/1000
489/489 [==============================] - 0s 137us/step - loss: 0.3524 - acc: 0.8200 - val_loss: 0.7290 - val_acc: 0.7012

5.3 分析模型

model.evaluate(x_train, y_train)
489/489 [==============================] - 0s 41us/step
[0.3090217998422728, 0.8548057079315186]
model.evaluate(x_test, y_test)
164/164 [==============================] - 0s 64us/step
[0.7289713301309725, 0.7012194991111755]
plt.plot(history.epoch, history.history.get('val_acc'), c='r', label='val_acc')
plt.plot(history.epoch, history.history.get('acc'), c='b', label='acc')
plt.legend()

【深度学习】实验13 使用Dropout抑制过拟合_第3张图片

6. 正则化

l1:loss = s*abs(w1 + w2 + …) + mse

l2:loss = s*(w12 + w22 + …) + mse

from keras import regularizers

6.1 神经网络太过复杂容易过拟合

#神经网络太过复杂容易过拟合
model = keras.Sequential()
model.add(layers.Dense(128, kernel_regularizer=regularizers.l2(0.001), input_dim=15, activation='relu'))
model.add(layers.Dense(128, kernel_regularizer=regularizers.l2(0.001), activation='relu'))
model.add(layers.Dense(128, kernel_regularizer=regularizers.l2(0.001), activation='relu'))
model.add(layers.Dense(1, activation='sigmoid'))
model.compile(optimizer='adam',
              loss='binary_crossentropy',
              metrics=['acc']
)
history = model.fit(x_train, y_train, epochs=1000, validation_data=(x_test, y_test))
Train on 489 samples, validate on 164 samples
Epoch 1/1000
489/489 [==============================] - 0s 752us/step - loss: 22.2560 - acc: 0.5910 - val_loss: 9.1111 - val_acc: 0.6524
Epoch 2/1000
489/489 [==============================] - 0s 137us/step - loss: 6.8963 - acc: 0.6217 - val_loss: 3.2886 - val_acc: 0.4573
Epoch 3/1000
489/489 [==============================] - 0s 161us/step - loss: 5.0407 - acc: 0.6830 - val_loss: 1.1973 - val_acc: 0.7256
Epoch 4/1000
489/489 [==============================] - 0s 218us/step - loss: 6.6088 - acc: 0.6421 - val_loss: 7.4651 - val_acc: 0.7012
Epoch 5/1000
489/489 [==============================] - 0s 233us/step - loss: 8.3945 - acc: 0.6973 - val_loss: 2.5579 - val_acc: 0.7317
Epoch 6/1000
489/489 [==============================] - 0s 192us/step - loss: 7.0204 - acc: 0.6196 - val_loss: 3.6758 - val_acc: 0.6829
Epoch 7/1000
489/489 [==============================] - 0s 152us/step - loss: 3.9961 - acc: 0.7382 - val_loss: 1.6183 - val_acc: 0.7317
Epoch 8/1000
489/489 [==============================] - 0s 94us/step - loss: 2.3441 - acc: 0.6237 - val_loss: 1.1523 - val_acc: 0.7256
Epoch 9/1000
489/489 [==============================] - 0s 114us/step - loss: 1.8178 - acc: 0.6442 - val_loss: 1.3449 - val_acc: 0.7073
Epoch 10/1000
489/489 [==============================] - 0s 157us/step - loss: 1.6122 - acc: 0.7117 - val_loss: 1.2869 - val_acc: 0.6646
……
Epoch 1000/1000
489/489 [==============================] - 0s 130us/step - loss: 0.1452 - acc: 0.9775 - val_loss: 1.0515 - val_acc: 0.7866
model.evaluate(x_train, y_train)
model.evaluate(x_train, y_train)
489/489 [==============================] - 0s 34us/step
[0.17742264538942426, 0.9611452221870422]
model.evaluate(x_test, y_test)
164/164 [==============================] - 0s 77us/step
[1.0514701096023, 0.7865853905677795]

6.2 太简单容易欠拟合

#太简单容易欠拟合
model = keras.Sequential()
model.add(layers.Dense(4, input_dim=15, activation='relu'))
model.add(layers.Dense(1,  activation='relu'))
model.add(layers.Dense(1, activation='sigmoid'))
model.compile(optimizer='adam',
              loss='binary_crossentropy',
              metrics=['acc']
)
history = model.fit(x_train, y_train, epochs=1000, validation_data=(x_test, y_test))
Train on 489 samples, validate on 164 samples
Epoch 1/1000
489/489 [==============================] - 0s 502us/step - loss: 0.6932 - acc: 0.4765 - val_loss: 0.6931 - val_acc: 0.6341
Epoch 2/1000
489/489 [==============================] - 0s 91us/step - loss: 0.6931 - acc: 0.5174 - val_loss: 0.6930 - val_acc: 0.6341
Epoch 3/1000
489/489 [==============================] - 0s 107us/step - loss: 0.6931 - acc: 0.5174 - val_loss: 0.6924 - val_acc: 0.6341
Epoch 4/1000
489/489 [==============================] - 0s 91us/step - loss: 0.6930 - acc: 0.5174 - val_loss: 0.6916 - val_acc: 0.6341
Epoch 5/1000
489/489 [==============================] - 0s 101us/step - loss: 0.6930 - acc: 0.5174 - val_loss: 0.6914 - val_acc: 0.6341
Epoch 6/1000
489/489 [==============================] - 0s 113us/step - loss: 0.6930 - acc: 0.5174 - val_loss: 0.6914 - val_acc: 0.6341
Epoch 7/1000
489/489 [==============================] - 0s 147us/step - loss: 0.6929 - acc: 0.5174 - val_loss: 0.6908 - val_acc: 0.6341
Epoch 8/1000
489/489 [==============================] - 0s 166us/step - loss: 0.6929 - acc: 0.5174 - val_loss: 0.6905 - val_acc: 0.6341
Epoch 9/1000
489/489 [==============================] - 0s 162us/step - loss: 0.6929 - acc: 0.5174 - val_loss: 0.6904 - val_acc: 0.6341
Epoch 10/1000
489/489 [==============================] - 0s 129us/step - loss: 0.6928 - acc: 0.5174 - val_loss: 0.6901 - val_acc: 0.6341
……
Epoch 1000/1000
489/489 [==============================] - 0s 86us/step - loss: 0.6926 - acc: 0.5174 - val_loss: 0.6849 - val_acc: 0.6341
model.evaluate(x_train, y_train)
489/489 [==============================] - 0s 43us/step
[0.6925447341854587, 0.5173823833465576]
model.evaluate(x_test, y_test)
164/164 [==============================] - 0s 39us/step
[0.684889389247429, 0.6341463327407837]

6.3 选取适当的神经网络

# 选取适当的神经网络
model = keras.Sequential()
model.add(layers.Dense(4, input_dim=15, activation='relu'))
model.add(layers.Dense(4,  activation='relu'))
model.add(layers.Dense(1, activation='sigmoid'))
model.compile(optimizer='adam',
              loss='binary_crossentropy',
              metrics=['acc']
)
history = model.fit(x_train, y_train, epochs=1000, validation_data=(x_test, y_test))
Train on 489 samples, validate on 164 samples
Epoch 1/1000
489/489 [==============================] - 0s 575us/step - loss: 40.1825 - acc: 0.5317 - val_loss: 17.6376 - val_acc: 0.6098
Epoch 2/1000
489/489 [==============================] - 0s 104us/step - loss: 30.0785 - acc: 0.5337 - val_loss: 12.6986 - val_acc: 0.6159
Epoch 3/1000
489/489 [==============================] - 0s 148us/step - loss: 20.0469 - acc: 0.5112 - val_loss: 8.3732 - val_acc: 0.5671
Epoch 4/1000
489/489 [==============================] - 0s 151us/step - loss: 12.5171 - acc: 0.4908 - val_loss: 3.8925 - val_acc: 0.5061
Epoch 5/1000
489/489 [==============================] - 0s 113us/step - loss: 4.4324 - acc: 0.4294 - val_loss: 0.9156 - val_acc: 0.4573
Epoch 6/1000
489/489 [==============================] - 0s 79us/step - loss: 1.0313 - acc: 0.5419 - val_loss: 0.9974 - val_acc: 0.4695
Epoch 7/1000
489/489 [==============================] - 0s 88us/step - loss: 1.0071 - acc: 0.5562 - val_loss: 0.8852 - val_acc: 0.5183
Epoch 8/1000
489/489 [==============================] - 0s 88us/step - loss: 0.9085 - acc: 0.5808 - val_loss: 0.7934 - val_acc: 0.5366
Epoch 9/1000
489/489 [==============================] - 0s 107us/step - loss: 0.8235 - acc: 0.5992 - val_loss: 0.7390 - val_acc: 0.5366
Epoch 10/1000
489/489 [==============================] - 0s 114us/step - loss: 0.7711 - acc: 0.5971 - val_loss: 0.7174 - val_acc: 0.5366
……
Epoch 1000/1000
489/489 [==============================] - 0s 141us/step - loss: 0.3095 - acc: 0.8732 - val_loss: 0.3971 - val_acc: 0.8537
model.evaluate(x_train, y_train)
489/489 [==============================] - 0s 68us/step
[0.30120014958464536, 0.8813905715942383]
model.evaluate(x_test, y_test)
164/164 [==============================] - 0s 45us/step
[0.39714593858253666, 0.8536585569381714]

附:系列文章

序号 文章目录 直达链接
1 波士顿房价预测 https://want595.blog.csdn.net/article/details/132181950
2 鸢尾花数据集分析 https://want595.blog.csdn.net/article/details/132182057
3 特征处理 https://want595.blog.csdn.net/article/details/132182165
4 交叉验证 https://want595.blog.csdn.net/article/details/132182238
5 构造神经网络示例 https://want595.blog.csdn.net/article/details/132182341
6 使用TensorFlow完成线性回归 https://want595.blog.csdn.net/article/details/132182417
7 使用TensorFlow完成逻辑回归 https://want595.blog.csdn.net/article/details/132182496
8 TensorBoard案例 https://want595.blog.csdn.net/article/details/132182584
9 使用Keras完成线性回归 https://want595.blog.csdn.net/article/details/132182723
10 使用Keras完成逻辑回归 https://want595.blog.csdn.net/article/details/132182795
11 使用Keras预训练模型完成猫狗识别 https://want595.blog.csdn.net/article/details/132243928
12 使用PyTorch训练模型 https://want595.blog.csdn.net/article/details/132243989
13 使用Dropout抑制过拟合 https://want595.blog.csdn.net/article/details/132244111
14 使用CNN完成MNIST手写体识别(TensorFlow) https://want595.blog.csdn.net/article/details/132244499
15 使用CNN完成MNIST手写体识别(Keras) https://want595.blog.csdn.net/article/details/132244552
16 使用CNN完成MNIST手写体识别(PyTorch) https://want595.blog.csdn.net/article/details/132244641
17 使用GAN生成手写数字样本 https://want595.blog.csdn.net/article/details/132244764
18 自然语言处理 https://want595.blog.csdn.net/article/details/132276591

你可能感兴趣的:(《,深度学习,》,python,深度学习,开发语言)