Experience-weighted attraction learning in normal form games
Experience-weighted attraction learning in normal form games
复制标题
DOI：
10.1111/1468-0262.00054
复制
发表时间：
1999-07-01
期刊：
影响因子：
6.1
通讯作者：
Ho, TH
中科院分区：
文献类型：
--
作者：
Camerer, C;Ho, TH
在“经验加权吸引力”(EWA)学习中，策略具有反映初始倾向的吸引力，根据回报经验进行更新，并根据某种规则(如Logit)确定选择概率。一个关键特征是一个参数Delta，它对没有根据收益选择的策略的假设强化的强度进行加权，相对于根据收到的收益对选择的策略进行强化。其他关键功能是两个折扣率，Phi和Rho，分别对以前的景点进行折扣，以及体验权重。EWA将强化学习和加权虚拟游戏(信念学习)作为特例，并混合了它们的关键要素。当增量=0和Rho=0时，累积选择加固结果。当增量=1，Rho=Phi时，在给定加权虚拟游戏信念的情况下，策略的强化水平与预期收益完全相同。使用三组实验数据，对模型的参数估计进行了部分数据校准，并用于预测一个坚守的样本。对德尔塔的估计一般都在附近。50，Phi在附近。8-1，Rho在0到Phi之间变化。强化和信念学习的特殊情况通常被拒绝，而偏向于Ewa，尽管信念模型在一些常和博弈中表现得更好。EWA能够结合以前方法的最佳特征，允许景点像选择强化那样灵活地开始和发展，但实质上像基于信念的模型隐含的那样强化未选择的策略。
In 'experience-weighted attraction' (EWA) learning, strategies have attractions that reflect initial predispositions, are updated based on payoff experience, and determine choice probabilities according to some rule (e.g., logit). A key feature is a parameter delta that weights the strength of hypothetical reinforcement of strategies that were not chosen according to the payoff they would have yielded, relative to reinforcement of chosen strategies according to received payoffs. The other key features are two discount rates, phi and rho, which separately discount previous attractions, and an experience weight. EWA includes reinforcement learning and weighted fictitious play (belief learning) as special cases, and hybridizes their key elements. When delta = 0 and rho = 0, cumulative choice reinforcement results. When delta = 1 and rho = phi, levels of reinforcement of strategies are exactly the same as expected payoffs given weighted fictitious play beliefs. Using three sets of experimental data, parameter estimates of the model were calibrated on part of the data and used to predict a holdout sample. Estimates of delta are generally around. 50, phi around. 8-1, and rho varies from 0 to phi. Reinforcement and belief-learning special cases are generally rejected in favor of EWA, though belief models do better in some constant-sum games. EWA is able to combine the best features of previous approaches, allowing attractions to begin and grow flexibly as choice reinforcement does, but reinforcing unchosen strategies substantially as belief-based models implicitly do.