marginal.sql_in File Reference

SQL functions for linear regression. More...

Functions
float8[]	marginal_logregr_step_transition (float8[] state, boolean y, float8[] x, float8[] coef)
float8[]	marginal_logregr_step_merge_states (float8[] state1, float8[] state2)
marginal_logregr_result	marginal_logregr_step_final (float8[] state)
aggregate marginal_logregr_result	marginal_logregr (boolean dependentVariable, float8[] independentVariables, float8[] coef)
	Compute marginal effects for logistic regression. More...
varchar	margins_logregr (varchar usage_string)
varchar	margins_logregr ()
void	margins_logregr (varchar source_table, varchar out_table, varchar dependent_varname, varchar input_independent_varname, varchar input_group_cols, integer[] marginal_vars, integer max_iter, varchar optimizer, float8 tolerance, boolean verbose_mode)
	A wrapper function for the various marginal regression analyzes. More...
CREATE OR REPLACE FUNCTION MADlib	margins_logregr (source_table VARCHAR--name of input table, out_table VARCHAR--name of output table, dependent_variable VARCHAR--name of dependent variable, independent_variable VARCHAR--name of independent variable) RETURNS VOID AS $$BEGIN PERFORM MADlib.margins_logregr(source_table
	Marginal effects with default variables. More...
CREATE OR REPLACE FUNCTION MADlib	margins_logregr (source_table VARCHAR--name of input table, out_table VARCHAR--name of output table, dependent_variable VARCHAR--name of dependent variable, independent_variable VARCHAR--name of independent variable, grouping_cols VARCHAR--name of grouping cols) RETURNS VOID AS $$BEGIN PERFORM MADlib.margins_logregr(source_table
	Marginal effects with default variable_names. More...
CREATE OR REPLACE FUNCTION MADlib	margins_logregr (source_table VARCHAR--name of input table, out_table VARCHAR--name of output table, dependent_variable VARCHAR--name of dependent variable, independent_variable VARCHAR--name of independent variable, grouping_cols VARCHAR--name of grouping cols, marginal_vars INTEGER[]--indices of variables to calculate marginal effects on) RETURNS VOID AS $$BEGIN PERFORM MADlib.margins_logregr(source_table
	Marginal effects with default variable_names. More...
CREATE OR REPLACE FUNCTION MADlib	margins_logregr (source_table VARCHAR--name of input table, out_table VARCHAR--name of output table, dependent_variable VARCHAR--name of dependent variable, independent_variable VARCHAR--name of independent variable, grouping_cols VARCHAR--name of grouping cols, marginal_vars INTEGER[]--indices of variables to calculate marginal effects on, max_iter INTEGER--Max iterations for the logstic regression inner call) RETURNS VOID AS $$BEGIN PERFORM MADlib.margins_logregr(source_table
	Marginal effects with default variable_names. More...
CREATE OR REPLACE FUNCTION MADlib	margins_logregr (source_table VARCHAR--name of input table, out_table VARCHAR--name of output table, dependent_variable VARCHAR--name of dependent variable, independent_variable VARCHAR--name of independent variable, grouping_cols VARCHAR--name of grouping cols, marginal_vars INTEGER[]--indices of variables to calculate marginal effects on, max_iter INTEGER--Max iterations for the logstic regression inner call, optimizer VARCHAR--Logistic regression optimizer) RETURNS VOID AS $$BEGIN PERFORM MADlib.margins_logregr(source_table
	Marginal effects with default variable_names. More...
CREATE OR REPLACE FUNCTION MADlib	margins_logregr (source_table VARCHAR--name of input table, out_table VARCHAR--name of output table, dependent_variable VARCHAR--name of dependent variable, independent_variable VARCHAR--name of independent variable, grouping_cols VARCHAR--name of grouping cols, marginal_vars INTEGER[]--indices of variables to calculate marginal effects on, max_iter INTEGER--Max iterations for the logstic regression inner call, optimizer VARCHAR--Logistic regression optimizer, tolerance DOUBLE PRECISION--Tolerance) RETURNS VOID AS $$BEGIN PERFORM MADlib.margins_logregr(source_table
	Marginal effects with default variable_names. More...
End of Default Variable calls for margins_logregr Marginal Multi Logistic Regression CREATE TYPE MADlib marginal_mlogregr_result	AS (margins DOUBLE PRECISION[], coef DOUBLE PRECISION[], std_err DOUBLE PRECISION[], t_stats DOUBLE PRECISION[], p_values DOUBLE PRECISION[])
CREATE OR REPLACE FUNCTION MADlib	mlogregr_marginal_step_transition (state DOUBLE PRECISION[], y INTEGER, num_categories INTEGER, ref_category INTEGER, x DOUBLE PRECISION[], coef DOUBLE PRECISION[]) RETURNS DOUBLE PRECISION[] AS 'MODULE_PATHNAME'LANGUAGE C IMMUTABLE
CREATE OR REPLACE FUNCTION MADlib	mlogregr_marginal_step_merge_states (state1 DOUBLE PRECISION[], state2 DOUBLE PRECISION[]) RETURNS DOUBLE PRECISION[] AS 'MODULE_PATHNAME'LANGUAGE C IMMUTABLE STRICT
CREATE OR REPLACE FUNCTION MADlib	mlogregr_marginal_step_final (state DOUBLE PRECISION[]) RETURNS MADlib.marginal_mlogregr_result AS 'MODULE_PATHNAME'LANGUAGE C IMMUTABLE STRICT

Variables
CREATE OR REPLACE FUNCTION MADlib	out_table
CREATE OR REPLACE FUNCTION MADlib	dependent_variable
CREATE OR REPLACE FUNCTION MADlib	independent_variable
CREATE OR REPLACE FUNCTION MADlib	NULL
CREATE OR REPLACE FUNCTION MADlib	irls
CREATE OR REPLACE FUNCTION MADlib	False
	END
LANGUAGE plpgsql	VOLATILE
CREATE OR REPLACE FUNCTION MADlib	grouping_cols
CREATE OR REPLACE FUNCTION MADlib	marginal_vars
CREATE OR REPLACE FUNCTION MADlib	max_iter
CREATE OR REPLACE FUNCTION MADlib	optimizer
CREATE OR REPLACE FUNCTION MADlib	tolerance

Detailed Description

Date

January 2011

See Also

Calculates marginal effects for various regression models.

Function Documentation

End of Default Variable calls for margins_logregr Marginal Multi Logistic Regression CREATE TYPE MADlib marginal_mlogregr_result AS	(	margins DOUBLE	PRECISION[],
		coef DOUBLE	PRECISION[],
		std_err DOUBLE	PRECISION[],
		t_stats DOUBLE	PRECISION[],
		p_values DOUBLE	PRECISION[]
	)

aggregate marginal_logregr_result marginal_logregr	(	boolean	dependentVariable,
		float8[]	independentVariables,
		float8[]	coef
	)

Parameters

dependentVariable	Column containing the dependent variable
independentVariables	Column containing the array of independent variables
coef	Column containing the array of the coefficients (as obtained by logregr)

To include an intercept in the model, set one coordinate in the independentVariables array to 1.

Returns

A composite value:

• margins FLOAT8[] - Array of marginal effects
• coefFLOAT8[] - The coefficients for the regression
• std_err FLOAT8[] - Array of standard-errors (calculated by the delta method),
• t_stats FLOAT8[] - Array of t-statistics
• p_values FLOAT8[] - Array of p-values

Usage

• Get all the diagnostic statistics:
  

 SELECT marginal_logregr(dependentVariable,
 independentVariables, coef)
 FROM dataTable;

marginal_logregr_result marginal_logregr_step_final

(

float8[]

state)

float8 [] marginal_logregr_step_merge_states	(	float8[]	state1,
		float8[]	state2
	)

float8 [] marginal_logregr_step_transition	(	float8[]	state,
		boolean	y,
		float8[]	x,
		float8[]	coef
	)

varchar margins_logregr

(

varchar

usage_string)

varchar margins_logregr

(

)

void margins_logregr	(	varchar	source_table,
		varchar	out_table,
		varchar	dependent_varname,
		varchar	input_independent_varname,
		varchar	input_group_cols,
		integer[]	marginal_vars,
		integer	max_iter,
		varchar	optimizer,
		float8	tolerance,
		boolean	verbose_mode
	)

Parameters

source_table	String identifying the input table
out_table	String identifying the output table to be created
dependent_varname	Column containing the dependent variable
input_independent_varname	Column containing the array of independent variables
grouping_cols	Set of columns to group by.
marginal_vars	Subset of independent variables to calculate marginal effects for.
max_iter	Maximum number of iterations
optimzer	Optimizer to be used (newton/irls, cg or idg)
tolerance	Resiual tolerance
verbose_mode	Verbose mode (on/off)

To include an intercept in the model, set one coordinate in the independentVariables array to 1.

Returns

void

Usage

For function summary information. Run sql> select margins_logregr('help'); OR sql> select margins_logregr(); OR sql> select margins_logregr('?'); For function usage information. Run sql> select margins_logregr('usage');

• Compute the coefficients, and the get the marginal diagnostic statistics:

    select margins_logregr(source_table, out_table, dependentVariable, independentVariables);
   

CREATE OR REPLACE FUNCTION MADlib margins_logregr	(	source_table VARCHAR--name of input	table,
		out_table VARCHAR--name of output	table,
		dependent_variable VARCHAR--name of dependent	variable,
		independent_variable VARCHAR--name of independent	variable
	)

CREATE OR REPLACE FUNCTION MADlib margins_logregr	(	source_table VARCHAR--name of input	table,
		out_table VARCHAR--name of output	table,
		dependent_variable VARCHAR--name of dependent	variable,
		independent_variable VARCHAR--name of independent	variable,
		grouping_cols VARCHAR--name of grouping	cols
	)

CREATE OR REPLACE FUNCTION MADlib margins_logregr	(	source_table VARCHAR--name of input	table,
		out_table VARCHAR--name of output	table,
		dependent_variable VARCHAR--name of dependent	variable,
		independent_variable VARCHAR--name of independent	variable,
		grouping_cols VARCHAR--name of grouping	cols,
		marginal_vars INTEGER--[]indices of variables to calculate marginal effects	on
	)

CREATE OR REPLACE FUNCTION MADlib margins_logregr	(	source_table VARCHAR--name of input	table,
		out_table VARCHAR--name of output	table,
		dependent_variable VARCHAR--name of dependent	variable,
		independent_variable VARCHAR--name of independent	variable,
		grouping_cols VARCHAR--name of grouping	cols,
		marginal_vars INTEGER--[]indices of variables to calculate marginal effects	on,
		max_iter INTEGER--Max iterations for the logstic regression inner	call
	)

CREATE OR REPLACE FUNCTION MADlib margins_logregr	(	source_table VARCHAR--name of input	table,
		out_table VARCHAR--name of output	table,
		dependent_variable VARCHAR--name of dependent	variable,
		independent_variable VARCHAR--name of independent	variable,
		grouping_cols VARCHAR--name of grouping	cols,
		marginal_vars INTEGER--[]indices of variables to calculate marginal effects	on,
		max_iter INTEGER--Max iterations for the logstic regression inner	call,
		optimizer VARCHAR--Logistic regression	optimizer
	)

CREATE OR REPLACE FUNCTION MADlib margins_logregr	(	source_table VARCHAR--name of input	table,
		out_table VARCHAR--name of output	table,
		dependent_variable VARCHAR--name of dependent	variable,
		independent_variable VARCHAR--name of independent	variable,
		grouping_cols VARCHAR--name of grouping	cols,
		marginal_vars INTEGER--[]indices of variables to calculate marginal effects	on,
		max_iter INTEGER--Max iterations for the logstic regression inner	call,
		optimizer VARCHAR--Logistic regression	optimizer,
		tolerance DOUBLE PRECISION--	Tolerance
	)

CREATE OR REPLACE FUNCTION MADlib mlogregr_marginal_step_final

(

state DOUBLE

PRECISION[])

CREATE OR REPLACE FUNCTION MADlib mlogregr_marginal_step_merge_states	(	state1 DOUBLE	PRECISION[],
		state2 DOUBLE	PRECISION[]
	)

CREATE OR REPLACE FUNCTION MADlib mlogregr_marginal_step_transition	(	state DOUBLE	PRECISION[],
		y	INTEGER,
		num_categories	INTEGER,
		ref_category	INTEGER,
		x DOUBLE	PRECISION[],
		coef DOUBLE	PRECISION[]
	)

Variable Documentation

CREATE OR REPLACE FUNCTION MADlib dependent_variable

END

CREATE OR REPLACE FUNCTION MADlib False

CREATE OR REPLACE FUNCTION MADlib grouping_cols

CREATE OR REPLACE FUNCTION MADlib independent_variable

CREATE OR REPLACE FUNCTION MADlib irls

CREATE OR REPLACE FUNCTION MADlib marginal_vars

CREATE OR REPLACE FUNCTION MADlib max_iter

CREATE OR REPLACE FUNCTION MADlib NULL

CREATE OR REPLACE FUNCTION MADlib optimizer

CREATE OR REPLACE FUNCTION MADlib out_table

CREATE OR REPLACE FUNCTION MADlib tolerance

LANGUAGE plpgsql VOLATILE