Q: How to implement StandardScalar? Do we have to? http://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.StandardScaler.html
Features | Feature Columns | Examples |
---|---|---|
Integer scalar | numeric_column('age') | 24 |
Float scaler | numeric_column('price') | $3.14 |
Normalized float | normalizer_fn=lambda e: tf.log(e + 1.0) | norm' price |
Features | Feature Columns | Examples |
---|---|---|
String linear | categorical_column_with_hash_bucket(100) | 'red' |
String linear | categorical_column_with_vocabulary_list | 'red' |
String deep | embedding_column(categorical_column, dim) | 'magenta' |
String deep list | embedding_column(categorical_column, dim) | groceries |
String deep 1 hot | indicator_column(categorical_column) | 0-1-0-0-0-0 |
Integerized/id | categorical_column_with_identity(ubound) | product_id |
Buketized intgr | bucketized_column(numeric_column) | age_bands |
Buketized float | bucketized_column(numeric_column) | price_bands |
Features | Feature Columns | Examples |
---|---|---|
Bucket x category | crossed_column(age_bands_col, color_col) | senior_grey |
Feature x magnitude | weighted_categorical_column(sku, weight) | a 9kg;z 5kg |