Stochastic models for learning.

Abstract

"The object of this book is to contribute to the study of learning--one of the many fields of psychology. A possible probabilistic framework or model for analyzing data from a variety of experiments on animal and human learning is presented. This model is described in general terms and is applied to a number of experiments. The stress is on the viewpoint that learning is probabilistic or stochastic--that events change the probabilities of the various responses. The book is divided into two main parts. The first presents the general model, describes many of its formal properties, and considers a number of special cases. The second part applies the model to a number of experiments on learning and treats the statistical problems of estimating parameters and measuring goodness of fit. The mathematics used is given in detail in an elementary manner. Serious mathematical difficulties are avoided by Monte Carlo or stat-rat trials that simulate learning experiments by means of routines based on random numbers. Stat-rat trials facilitate the comparison of actual observations with the model. Several special tables will help the reader apply the model to his own experiments"--Jacket. (PsycINFO Database Record (c) 2014 APA, all rights reserved)