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

Question

In Tensorflow there is already a function to create feature by crossing columns tf.feature_column.crossed_column , but it is more for category data. How about numeric data?

For example, there are 2 columns already

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

if i want to create a third and fourth feature columns base on age and education_num like this

my_feature = age * education_num
my_another_feature = age * age

How can it be done?

Answer 1

You can declare a custom numerical column and add it to the dataframe in your input function:

# 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())