电影评论：二分类

跟着《python深度学习》做了个书上小项目，现在总结下该项目内容。

1. 获取数据(这里获取的是IMDB数据集，其中有train_data和train_label两个标签，train_data：英文句子，train_label：正/负面(0,1))

   from keras.datasets import imdb
   (train_data, train_labels), (test_data, test_labels) = imdb.load_data(num_words=10000)
   

2. 数据预处理(要将列表转化为张量，一般采用填充列表或one-hot编码方式，填充列表是将列表拥有相同的长度，ont-hot编码是将数据转化为只有0和1)

   import numpy as np
   
   def vectorize_sequences(sequences, dimension=10000):
       results = np.zeros((len(sequences), dimension))
       for i, sequence in enumerate(sequences):
           results[i, sequence] = 1
       return results
   
   x_train = vectorize_sequences(train_data)
   x_test = vectorize_sequences(test_data)
   
   # 或者标签向量化采用下面方式
   y_train = np.asarray(train_labels).astype('float32')
   y_test = np.asarray(train_labels).astype('float32')
   

3. 构建网络(这里采用两层relu层(units=16)+一层sigmoid层)

   from keras import models
   from keras import layers
   
   model = models.Sequential()
   model.add(layers.Dense(16, activation='relu', input_shape=(10000, )))
   model.add(layers.Dense(16, activation='relu'))
   model.add(layers.Dense(1, activation='sigmoid'))
   

4. 编译模型(采用二元交叉熵损失函数(binary_crossentropy))

   model.compile(optimizer='rmsprop',
                 loss='binary_crossentropy',
                 metrics=['accuracy'])
   

5. 验证模型(留出10000数据作为验证，其余进行训练，训练后发现第四轮开始验证损失和精度越来越不好，说明开始过拟合)

   # 留出验证集
   x_val = x_train[:10000]
   partial_x_train = x_train[10000:]
   
   y_val = y_train[:10000]
   partial_y_train = y_train[10000:]
   
   
   # 训练模型
   model.compile(optimizer='rmsprop',
                 loss='binary_crossentropy',
                 metrics=['acc'])
   
   history = model.fit(partial_x_train,
                       partial_y_train,
                       epochs=20,
                       batch_size=512,
                       validation_data=(x_val, y_val))
   

6. 重新训练网络，设置训练轮数为4轮。

   model = models.Sequential()
   model.add(layers.Dense(16, activation='relu', input_shape=(10000,)))
   model.add(layers.Dense(16, activation='relu'))
   model.add(layers.Dense(1, activation='sigmoid'))
   
   model.compile(optimizer='rmsprop',
                 loss='binary_crossentropy',
                 metrics=['accuracy'])
   
   model.fit(x_train, y_train, epochs=4, batch_size=512)
   result = model.evaluate(x_test, y_test)
   
   print(result)