如何通过乘以另外两个 tf.feature_column 来创建 tf.feature_column？

Question

Tensorflow中已经有跨列创建特征的功能tf.feature_column.crossed_column，但更多的是针对类别数据。数字数据呢？

例如，已经有2列

age = tf.feature_column.numeric_column("age")
education_num = tf.feature_column.numeric_column("education_num")

如果我想根据年龄和 education_num 创建第三个和第四个特征列，就像这样

my_feature = age * education_num
my_another_feature = age * age

如何做到？

Answer 1

您可以声明自定义数值列并将其添加到 input function 中的数据框：

# Existing features
age = tf.feature_column.numeric_column("age")
education_num = tf.feature_column.numeric_column("education_num")
# Declare a custom column just like other columns
my_feature = tf.feature_column.numeric_column("my_feature")

...
# Add to the list of features
feature_columns = { ... age, education_num, my_feature, ... }

...
def input_fn():
  df_data = pd.read_csv("input.csv")
  df_data = df_data.dropna(how="any", axis=0)
  # Manually update the dataframe
  df_data["my_feature"] = df_data["age"] * df_data["education_num"]

  return tf.estimator.inputs.pandas_input_fn(x=df_data,
                                             y=labels,
                                             batch_size=100,
                                             num_epochs=10)

...
model.train(input_fn=input_fn())

如何通过乘以另外两个 tf.feature_column 来创建 tf.feature_column？

How to create a tf.feature_column by multiplying two other tf.feature_columns?

python

tensorflow

machine-learning

feature-extraction

tensorflow-estimator