为 2d 分类数据正确的 Conv1D 层形状

Question

我的数据集（我们进行二进制分类的网络流量数据集）-

X_train 和 y_train 的形状是 (45447, 25) (45447,)

我正在研究的 model -

# fit and evaluate a model
import tensorflow as tf
def evaluate_model(X_train, y_train,X_test,y_test):

  X_train = X_train.reshape(45447,25,1)
  y_train=y_train.reshape(45447,1)
  verbose=0
  epochs=10
  batch_size = 32
  n_timesteps = X_train.shape[0]
  n_features= X_train.shape[1]
  print(n_timesteps,n_features)
  n_outputs = 1
  model = Sequential()
  model.add(Conv1D(filters=64, kernel_size=3, activation='relu', input_shape=(n_timesteps,n_features)))
  model.add(Conv1D(filters=64, kernel_size=3, activation='relu'))
  model.add(Dropout(0.5))
  model.add(MaxPooling1D(pool_size=2))
  model.add(Flatten())
  model.add(Dense(100, activation='relu'))
  model.add(Dense(n_outputs, activation='softmax'))
  model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
  # fit network
  #train_data = tf.data.Dataset.from_tensor_slices((X_train, y_train))
  #valid_data = tf.data.Dataset.from_tensor_slices((X_test, y_test))

  model.fit(X_train, y_train,epochs=10, batch_size=32, verbose=0)
      # evaluate model
  #_, accuracy = model.evaluate(X_test, y_test, batch_size=batch_size, verbose=0)
  

# summarize scores
def summarize_results(scores):
    print(scores)
    m, s = mean(scores), std(scores)
    print('Accuracy: %.3f%% (+/-%.3f)' % (m, s))

# run an experiment
def run_experiment(repeats=10):
    # load data
    
    # repeat experiment
    scores = list()
    for r in range(repeats):
        score = evaluate_model(X_train, y_train,X_test,y_test)
        score = score * 100.0
        print('>#%d: %.3f' % (r+1, score))
        scores.append(score)
    # summarize results
    summarize_results(scores)

# run the experiment
run_experiment()

我尝试过的-

1）我尝试将 pandas 数据转换为 numpy

2）将二维阵列重新整形为 3d -

X_train = X_train.reshape(45447,25,1)
y_train=y_train.reshape(45447,1,1)

3. 将数据制成 tf 对象-

   train_data = tf.data.Dataset.from_tensor_slices((X_train, y_train)) valid_data = tf.data.Dataset.from_tensor_slices((X_test, y_test))

我仍然无法运行我的 model ..它不断给出形状错误。 请帮助我了解 model 的形状。

Answer 1

我认为您必须从 input_shape 中删除批次并添加 1 作为长度或特征：

model.add(Conv1D(filters=64, kernel_size=3, activation='relu', input_shape=(n_features , 1)))

对于 output 层，使用一个带有 softmax 激活的单元密集单元是没有意义的，将 output 单元的数量更改为：

 n_outputs = your_number_of_categories

Answer 2

input_shape应该是X_train.shape[1:] 。 由于标签以整数形式提供，因此损失 function 应该是sparse_categorical_accuracy 。

def evaluate_model(X_train, y_train,X_test,y_test):

  X_train = X_train.reshape(45447,25,1)
  y_train=y_train.reshape(45447,1)
  verbose=0
  epochs=10
  batch_size = 32
  n_outputs = 2
  model = Sequential()
  model.add(Conv1D(filters=64, kernel_size=3, activation='relu', input_shape=X_train.shape[1:]))
  model.add(Conv1D(filters=64, kernel_size=3, activation='relu'))
  model.add(Dropout(0.5))
  model.add(MaxPooling1D(pool_size=2))
  model.add(Flatten())
  model.add(Dense(100, activation='relu'))
  model.add(Dense(n_outputs, activation='softmax'))
  model.compile(loss='sparse_categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
  # fit network
  model.fit(X_train, y_train,epochs=10, batch_size=32, verbose=0)
  # evaluate model
  _, accuracy = model.evaluate(X_test, y_test, batch_size=batch_size, verbose=0)