Method and system for using cooperative game theory to resolve statistical joint effects
First Claim
1. A method for constructing a statistical cooperative game, comprising:
- identifying a set of players for a statistical cooperative game;
identifying an access relationship between coalitions of the statistical cooperative game and elements of a multivariate statistical model, wherein a selected subset of the identified players is a coalition; and
determining a worth for selected coalitions in the statistical cooperative game based on elements of the multivariate statistical model accessible by a coalition.
0 Assignments
0 Petitions
Accused Products
Abstract
A method and system for cooperative resolution of joint statistical effects. A statistical cooperative game is used to represent statistical joint effects in a multivariate statistical model. Access relationships are created between players in a cooperative game and variables in a multivariate statistical model. A worth of a coalition in a cooperative game is determined, based on a multivariate statistical model and a performance measure of the multivariate statistical model. Cooperative resolution methods are applied to particular analytical procedures. Thus, the present invention may be used to construct statistical cooperative games and use cooperative game theory to resolve statistical joint effects in a variety of situations. The methods may be applicable to other types of joint effects problems such as those found in engineering, finance and other disciplines.
30 Citations
41 Claims
-
1. A method for constructing a statistical cooperative game, comprising:
-
identifying a set of players for a statistical cooperative game;
identifying an access relationship between coalitions of the statistical cooperative game and elements of a multivariate statistical model, wherein a selected subset of the identified players is a coalition; and
determining a worth for selected coalitions in the statistical cooperative game based on elements of the multivariate statistical model accessible by a coalition. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16)
identifying a plurality of elements of the multivariate statistical model;
identifying a set of coalitions in the statistical cooperative game; and
specifying an access relationship comprising a set of rules, wherein the set of rules determine for selected coalitions in the identified set of coalitions, elements that are accessible by the coalition and how the accessible elements may be used by the coalition.
-
-
6. The method of claim 1 wherein the step of determining a worth further comprises:
-
selecting a performance measure of the multivariate statistical model;
computing a performance measure for selected collations in the statistical cooperative game based on the elements of the multivariate statistical model accessible by a selected coalition; and
determining, for selected coalitions, a worth of the coalition in the statistical cooperative game based on the computed performance measure for that coalition.
-
-
7. The method of claim 1 further comprising determining allocations to players of a statistical cooperative game using a cooperative allocation procedure.
-
8. The method of claim 7 wherein determined allocations to the players of the statistical cooperative game are used to resolve statistical joint effects in a multivariate statistical model.
-
9. The method of claim 7 wherein the step of determining allocations to the players of the statistical cooperative game includes identifying a single such allocation for a single player.
-
10. The method of claim 7 wherein the cooperative allocation procedure includes a Shapley value, a proportional value, a powerpoint, a weighted Shapley value, or a log-linear value.
-
11. The method of claim 1 further comprising:
- applying the steps of claim 1 in a recursive manner to allocate a value allocated to a player accessing a plurality of variables in a first statistical cooperative game on the basis of a second cooperative game embodying a second set of players.
-
12. The method of claim 1 wherein the multivariate statistical model includes continuous independent variables.
-
13. The method of claim 1 wherein the multivariate statistical model includes categorical independent variables.
-
14. The method of claim 1 wherein the multivariate statistical model includes frequency data to compute marginal frequencies.
-
15. The method of claim 1 wherein the multivariate statistical model includes an ordinary least squares model, a time series model, an analysis of categorical effects model, an analysis of changes in proportions model, a covariance matrix, a capital asset pricing model, an arbitrage pricing theory model, an options pricing model, a derivatives pricing model, a Sharpe style analysis model, a macroeconomic model, a price forecasting model, a sales forecasting model, or a basic or generalized Brinson and Falcher manager attribution model.
-
16. The method of claim 1, further comprising:
-
including as independent variables timing or selection factors of performance attribution dimensions; and
defining the worth of a coalition to include an incremental performance resulting from inclusion of those timing or selection factors of performance attribution dimensions accessible in a coalition.
-
-
17. A method for constructing an access relationship between coalitions in a statistical cooperative game and a multivariate statistical model, comprising:
-
identifying a plurality of elements of the multivariate statistical model;
identifying a set of coalitions in the statistical cooperative game; and
specifying an access relationship comprising a set of rules, wherein the set of rules determine for selected coalitions in the identified set of coalitions, elements that are accessible by the coalition and how the accessible elements may be used by the coalition. - View Dependent Claims (18, 19, 20, 21, 22, 23, 24, 25)
a dependent variable is a returns time series for a financial security;
independent variables include return time series for a set of asset class benchmarks;
an asset class benchmark has a primary access relationship with a single player;
a measure of explanatory power is a R2 coefficient;
submodels are constructed for sets of independent variables corresponding to all coalitions in the game;
ora proportional value of the dual game is used determine allocations to players.
-
-
26. A method for determining a worth for selected coalitions in a statistical cooperative game, comprising:
-
selecting a performance measure of a multivariate statistical model;
computing a performance measure for selected collations in the statistical cooperative game based on elements of the multivariate statistical model accessible by a selected coalition; and
determining, for each coalition from the selected set of coalitions, a worth of a coalition in the statistical cooperative game based on the computed performance measure for that coalition. - View Dependent Claims (27, 28, 29, 30, 31, 32, 33, 34, 35)
-
-
36. A method for allocating a worth of a coalition in a cooperative game on a multiplicative basis, comprising:
-
generating a second cooperative game from a first cooperative game by setting a worth of a plurality of coalitions in the second cooperative game to a logarithm of the worth of a same coalition plus a constant;
applying a cooperative allocation procedure to the second cooperative game; and
creating an allocation for a player in the first cooperative game from an allocation in the second cooperative game by applying an antilog to a value allocated to a player in the statistical cooperative game and subtracting a constant. - View Dependent Claims (37, 38)
wherein the summation is over all coalitions S that contain player i, n is a number of players in the cooperative game, and s i s a number of players in a coalition S.
-
-
39. A method to estimate a total effect of a basic variable on an individual observation, set of observations, or forecast value in a multivariate statistical model with interaction variables, comprising:
-
constructing a multivariate statistical model with basic variables and a plurality of interaction variables derived from the basic variables;
identifying an access relationship between coalitions in a statistical cooperative game and elements of the multivariate statistical model, including the basic and interaction independent variables;
constructing a cooperative game based on the access relationship; and
applying a value function or other allocation rule to the cooperative game to attribute all effects among players in the game. - View Dependent Claims (40)
-
-
41. A cooperative game resolution system, comprising in combination:
-
a plurality of software modules stored as data bits in memory on one or more computers with one or more processors, the plurality of software modules including;
a player module for identifying a set of players for a statistical cooperative game, an access module for identifying an access relationship between coalitions of the statistical cooperative game and elements of a multivariate statistical, wherein a selected subset of the identified players is a coalition, and a worth module for determining a worth for selected coalitions in the statistical cooperative game based on elements of the multivariate statistical model accessible by a coalition or for determining a worth of a coalition in a cooperative game on a multiplicative basis; and
one or more databases for storing cooperative game data.
-
Specification