Category Similarity And Presentation Of The Experimental Study Of The Impact Of The Category Learning

The models of similarity comparison in the study of categorization in psychology included the Spatial Models, the Feature-based model of comparison and the structural alignment view of similarity. People have found that the structural alignment is helpful in category learning and participants were inclined to learn category by the structural alignment by Inference-earning. We supposed that whether we could learn categories by the structural alignment process depended on how the items were represented and on how those items were placed in correspondence. And Inference-learning was more convenient for those rules to be used because of its own features. Therefore, people learn by inference could learn categories better than those learn by categorization.Based on the theory and application on the structural alignment, the current study focused on the influence of structural alignment and similarity on category learning when the representation of items and the way they were placed in correspondence were changed. Participants in the research consisted of 322 senior under-graduates in South China Normal University. The materials were schematic drawings of bugs including two different categories. Participants were arranged randomly into two comparable learning conditions: Classification-learning condition and Inference-learning condition. Participants in Classification-learning condition were supposed to learn the category by predicting the category label based on the features of each stimulus and those in Inference-learning learn by predicting the missing feature based on the category label and parts of features of each stimulus.Analyzed the proportion of category-accordance response, the results of the three experiments showed that: (1) although in the condition of increased features of stimulus, participants in Inference-learning could acquire the categories more easily than those in Classification-learning by structural alignment;(2) with the similarity decreased, participants in Inference-learning even lost their superiority until we modify the manifestation of the stimuli by pairs to let them learning low-similarity categories by structural alignment;(3)in manifestation by pairs of stimuli, participants in Classification-learning could acquire the categories as easily as those in Inference-learning. What's more, the less differences and more commonness between the pairs, the more easily they acquire the categories by structural alignment.