Shapley Value Functions (ML Engine) - Teradata Vantage

Machine Learning Engine Analytic Function Reference

Product
Teradata Vantage
Release Number
9.02
9.01
2.0
1.3
Published
February 2022
Language
English (United States)
Last Update
2022-02-10
dita:mapPath
rnn1580259159235.ditamap
dita:ditavalPath
ybt1582220416951.ditaval
dita:id
B700-4003
lifecycle
previous
Product Category
Teradata Vantageā„¢

The Shapley value is intended to reflect the importance of each player to the coalition in a cooperative game (a game between coalitions of players, rather than between individual players).

ML Engine Shapley Value feature computes an approximate Shapley value by sampling over coalitions whose observed payoff values are included in the user data.

Function Description
ShapleyGenerateCombination (ML Engine) Takes combinations of players (coalitions) and creates input for ShapleyAddOnePlayer.

Input is either unsorted user data or sorted output from nPath function. If input is unsorted, ShapleyGenerateCombination inputs it to ShapleySortCombinations.

ShapleySortCombinations (ML Engine) Sorts combinations of players.

Input is from either ShapleyGenerateCombination or user.

ShapleyAddOnePlayer (ML Engine) Takes sorted combinations and outputs a table.

Input is from either ShapleyGenerateCombination or ShapleySortCombinations.

The SQL Statements to Compute the Shapley Value query the ShapleyAddOnePlayer input and output tables.

Computing a Shapley Value
How Machine Learning Engine functions compute a Shapley value