I was going through "Wide & Deep Learning" tensorflow tutorial & it's quite simply explained the process. But I missed few of the things. If someone can please explain them to me, it will be of great help:
1) Why occupation
and native_country
we are using tf.feature_column.categorical_column_with_hash_bucket
and again using tf.feature_column.embedding_column
?
2) Why in tf.feature_column.embedding_column
, we are taking dimension=8
, even though they have more unique values?
3) Why we are using crossed_columns
variables? In the document there's an explaination given. But I'm not fully understanding it.
The questions might sound silly, but I'm not sure of the answers to these questions.
Thank you!