pygambit.qre.logit_estimate#

pygambit.qre.logit_estimate(data: MixedStrategyProfile | MixedBehaviorProfile, use_empirical: bool = False, local_max: bool = False, first_step: float = 0.03, max_accel: float = 1.1) → LogitQREMixedStrategyFitResult | LogitQREMixedBehaviorFitResult#

Use maximum likelihood estimation to find the logit quantal response equilibrium which best fits empirical frequencies of play.

Added in version 16.3.0.

Parameters:

data (MixedStrategyProfile or MixedBehaviorProfile) – The empirical distribution of play to which to fit the QRE. To obtain the correct resulting log-likelihood, these should be expressed as total counts of observations of each action rather than probabilities. If a MixedBehaviorProfile is specified, estimation is done using the agent QRE.
use_empirical (bool, default = False) – If specified and True, use the empirical payoff approach for estimation. This replaces the payoff matrix of the game with an approximation as a collection of individual decision problems based on the empirical expected payoffs to strategies or actions. This is computationally much faster but in most cases produces estimates which deviate systematically from those obtained by computing the QRE correspondence of the game. See the discussion in [1] for more details.
local_max (bool, default False) –
The default behavior is to find the global maximiser along the principal branch. If this parameter is set to True, tracing stops at the first interior local maximiser found.

Note

This argument only has an effect when use_empirical is False.
first_step (float, default .03) –
The arclength of the initial step.

Note

This argument only has an effect when use_empirical is False.
max_accel (float, default 1.1) –
The maximum rate at which to lengthen the arclength step size.

Note

This argument only has an effect when use_empirical is False.

Returns:

The result of the estimation represented as a LogitQREMixedStrategyFitResult or LogitQREMixedBehaviorFitResult object, as appropriate.

Return type:

References