Table | Description |
---|---|
AttributeTable | Contains attribute names and values. |
ResponseTable | Contains attribute names and values and response values. |
CategoricalAttributeTable | [Optional] Contains categorical attribute names. Every categorical attribute in AttributeTable must be in non-empty partition in this table. |
AttributeTable Schema
Column | Data Type | Description |
---|---|---|
id_column | INTEGER, SMALLINT, BIGINT, NUMERIC, VARCHAR, or VARBYTE(n) | Point identifier. Cannot be NULL. |
attribute_name_column | VARCHAR | [Column appears once for each attribute_name_column.] Attribute name. Cannot be NULL. |
attribute_value_column | Numeric attribute: INTEGER, BIGINT, NUMERIC, or DOUBLE PRECISION Categorical attribute: Any |
Attribute value. If NULL, function estimates value by arithmetic means on an attribute basis. If estimate is out of range, function cannot use it to partition training data, so it is useless. |
actual_label | VARCHAR | Actual label of data point. |
ResponseTable Schema
Column | Data Type | Description |
---|---|---|
id_column | INTEGER, SMALLINT, BIGINT, NUMERIC, VARCHAR, or VARBYTE(n) | Point identifier. Cannot be NULL. |
response_column | INTEGER, BIGINT, NUMERIC, VARCHAR, or DOUBLE PRECISION | Response value for point. Can be NULL. |
CategoricalAttributeTable Schema
Column | Data Type | Description |
---|---|---|
categorical_attribute_name_column | VARCHAR | Categorical attribute name. Every categorical attribute in AttributeTable must be in non-empty partition in this table. |