基本卷积神经网络（CNN）

-CNN的基本结构：CNN与MLP相似，因为它们只向前传送信号（前馈网络），但有CNN特有的不同类型的层

Convolutional layer：在一个小的感受野（即滤波器）中处理数据
Pooling layer：沿2维向下采样（通常为宽度和高度）
Dense (fully connected) layer：类似于MLP的隐藏层

import numpy as np
import matplotlib.pyplot as plt

from sklearn import datasets
from sklearn.model_selection import train_test_split
from keras.utils.np_utils import to_categorical

加载数据集

sklearn中的数字数据集
文档：http://scikit-learn.org/stable/auto_examples/datasets/plot_digits_last_image.html

data = datasets.load_digits()
plt.imshow(data.images[0])    # show first number in the dataset
plt.show()
print('label: ', data.target[0])    # label = '0'

X_data = data.images
y_data = data.target
# shape of data
print(X_data.shape)    # (8 X 8) format
print(y_data.shape)

# reshape X_data into 3-D format
X_data = X_data.reshape((X_data.shape[0], X_data.shape[1], X_data.shape[2], 1))
# one-hot encoding of y_data
y_data = to_categorical(y_data)

将数据划分为列车/测试集

X_train, X_test, y_train, y_test = train_test_split(X_data, y_data, test_size = 0.3, random_state = 777)
print(X_train.shape)
print(X_test.shape)
print(y_train.shape)
print(y_test.shape)

from keras.models import Sequential
from keras import optimizers
from keras.layers import Dense, Activation, Flatten, Conv2D, MaxPooling2D

1.创建模型

创建模型与MLP（顺序）相同

model = Sequential()

2.卷积层

通常，二维卷积层用于图像处理
滤波器的大小（由“kernel\u Size”参数指定）定义感受野的宽度和高度**
过滤器数量（由“过滤器”参数指定）等于下一层的深度
步幅（由“步幅”参数指定）是过滤器每次移动改变位置的距离
图像可以零填充以防止变得太小（由“padding”参数指定）
Doc: https://keras.io/layers/convolutional/

# convolution layer
model.add(Conv2D(input_shape = (X_data.shape[1], X_data.shape[2], X_data.shape[3]), filters = 10, kernel_size = (3,3), strides = (1,1), padding = 'valid'))

3. 激活层

与 MLP 中的激活层相同
一般情况下，也使用relu
Doc: http://cs231n.github.io/assets/cnn/depthcol.jpeg

model.add(Activation('relu'))

4. 池化层

一般使用最大池化方法
减少参数数量
文档：https://keras.io/layers/pooling/

model.add(MaxPooling2D(pool_size = (2,2)))

5. Dense（全连接层）

卷积和池化层可以连接到密集层
文档：https://keras.io/layers/core/

# prior layer should be flattend to be connected to dense layers
model.add(Flatten())
# dense layer with 50 neurons
model.add(Dense(50, activation = 'relu'))
# final layer with 10 neurons to classify the instances
model.add(Dense(10, activation = 'softmax'))

6. Model compile & train

adam = optimizers.Adam(lr = 0.001)
model.compile(loss = 'categorical_crossentropy', optimizer = adam, metrics = ['accuracy'])
history = model.fit(X_train, y_train, batch_size = 50, validation_split = 0.2, epochs = 100, verbose = 0)
plt.plot(history.history['acc'])
plt.plot(history.history['val_acc'])
plt.legend(['training', 'validation'], loc = 'upper left')
plt.show()

results = model.evaluate(X_test, y_test)
print('Test accuracy: ', results[1])

（完）