Rule-Plus-Exception Model of Classification Learning.

摘要: The authors propose a rule-plus-exception model (RULEX) of classification learning. According to RULEX, people learn classify objects by forming simple logical rules and remembering occasional exceptions those rules. Because the learning process in RULEX is stochastic, predicts that individual Ss will vary greatly particular are formed stored. Averaged data presumed represent mixtures these highly idiosyncratic exceptions. accounts for numerous fundamental phenomena, including prototype specific exemplar effects, sensitivity correlational information, difficulty linearly separable versus nonlinearly categories, selective attention concepts with differing complexity. also distributions generalization patterns observed at subject level.