Learning And Transfer Of Rule-Based And Similarity-Based Categories In Comparison Task

Category learning is a process of learning to judge the category members by practice. In daily life, people learn and use categories in a variety of ways, they use categorical information to make inferences about unobserved features, to make decision, and to problem solve. Nature is complex and volatile, we are faced with the uncertain classification structure in many cases. The category learning process is a by-product of the interaction of the external objects and us. we will be limited the further research of category learning if we overemphasize direct classification and inference learning, because the use of category knowledge is very rich and flexible in our day-to-day lives. Comparison is an important thinking process, also a kind of important thinking methods. We can find the differences, similarities and relationships by comparing the various things and phenomena. A large number of studies have demonstrated that compare-based learning can promote us to acquire abstract concepts. This research use comparison task for integrating direct category learning and indirect category learning on the basis of previous studies, and it will explore rule-based and similarity-based category learning and learning strategies in comparison task.This research use a paradigm which is called the "same"-"different" categorization task. The participants are shown two stimuli simultaneously to make the same or different judgment in learning phase, but they need to classify the old stimulus and transfer stimulus during the testing phase. This research includes two experiments. Experiment 1 explores whether the category knowledge can be acquired by using the "same"-"different" categorization task, and compares single dimension rule-based learning and transfer in direct classification task presented individually stimulus with comparison task. Experiment 2 explores to compare two dimension rule-based category learning and transfer with similarity-based category structure in comparison task.The two experiments lead to some preliminary conclusions:Firstly, people can acquire single dimension rule-based category knowledge in comparison task and transfer it to the classification task by a compare-based a paradigm which is called the "same"-"different" categorization task. This paradigm can be used to research category learning.Secondly, single dimension rule-based category structure can be better acquired in conparison task than the classification task.Thirdly, comparison learning is more conducive to the transfer of acquired single dimension rule-based category than the classification learning.Fourthly, similarity-based category structure group learning is superior to the two dimension rule-based category structures group in comparison task.Fifth, only the rule-based category structures can be transferred from the comparison task to the classification task, while the similarity-based category structures can’t be transferred from the comparison task to the classification task.Sixth, whether two dimension rule-based or similarity-based category structures, people tend to use either the rule-based categorization strategy or the similarity-based categorization strategy.