table_to_pmml.sql_in File Reference

Functions
xml	pmml (varchar model_table)
	Given the model constructed from a data mining algorithm, this function converts the model into PMML format. More...
xml	pmml (varchar model_table, varchar name_spec)
	Given the model constructed from a data mining algorithm, this function converts the model into PMML format. More...
xml	pmml (varchar model_table, varchar[] name_spec)
	Given the model constructed from a data mining algorithm, this function converts the model into PMML format. More...
text	pmml ()

Function Documentation

pmml() [1/4]

xml pmml

(

varchar

model_table

)

Parameters

model_table

Model table that contains the output of a training algorithm.

Returns

An XML document corresponding to the PMML of the input model.

The names in the Data Dictionary are generated based on the names used in the model table. If the fields are represented as an array of expressions without array concatenation, the names in the Data Dictionary will match the source table, except that a suffix '_pmml_prediction' is added to the predicted variable. If the field is an array of FLOAT8 values, the names in the Data Dictionary will be indexed using the array name, like 'ind[1]', 'ind[2]', etc. In all other cases, the names will simply be 'x[1]', 'x[2]', and so on.

pmml() [2/4]

xml pmml	(	varchar	model_table,
		varchar	name_spec
	)

Parameters

model_table	Model table that contains the output of a training algorithm.
name_spec	Names to be used in the Data Dictionary, given as a string containing an expression such as 'y ~ x1 + x2 + x3'.

Returns

An XML document corresponding to the PMML of the input model.

pmml() [3/4]

xml pmml	(	varchar	model_table,
		varchar []	name_spec
	)

Parameters

model_table	Model table that contains the output of a training algorithm.
name_spec	Names to be used in the Data Dictionary given as an array of strings, such as ARRAY['y', 'x1', 'x2', ...].

Returns

An XML document corresponding to the PMML of the input model.

pmml() [4/4]

text pmml

(

)