Optimal Procedure for an n-Stage Testing and Learning Process

摘要

The paper extends the main results of AD-610 700 ('Optimal Stimulus Presentation Strategy for a Stimulus Sampling Model of Learning - II,' SDC document SP-1898, December 1964) to a more general learning model than the one considered there. The new model allows the 'learning rate,' i. e., the probability of transition from the unconditioned to the conditioned state in a single trial, to be different for a subject who has responded incorrectly in a trial than for one who has responded correctly. These rates were previously assumed to be the same. (Author)