MARKOV LEARNING MODELS FOR MULTIPERSON SITUATIONS, THE THEORY

摘要

The present chapter describes the underlying theory used in our experiments. The second chapter is concerned with various methods, statistical and otherwise, used in the analysis of data. Because the fundamental theory is probabilistic in character, the conceptual separation between the first two chapters is not absolutely clear. Theoretical quantities which do not depend on observed quantities are derived in the first chapter;examples are asymptotic mean probabilities of response and associated variances. Quantities which do depend on observed data are derived in the second chapter;a typical instance is the derivation of the maximum likelihood estimate of the learning parameter.nChapter 3 is concerned with some simple zero-sun, two-person games, and Chapter 4 with some non-zero-sum, two-person games. Chapter 5 deals with the analysis of game-theoretical information from the standpoint of what learning theorists call discrimination theory. Chapter 6 considers experiments concerned with a three-person, simple majority game. Chapter 7 describes some experiments in which the subjects were told various things about the pay-off matrix. Chapter 8 analyzes the effects of monetary pay-offs.