A Connectionist Model of Categorization Response Times
Although perceptual categorisation has been studied extensively in psychology, response times in categorisation tasks have only recently become an important research topic (e.g., [1,2,3,4,5,6,7,8,9]). In this article, we propose a connectionist model of categorisation RT, called CONCAT, which aims to provide a joint account of response times and choice proportions in binary classification tasks. First, we outline the basic principles of the model. Next, we present a perceptual categorisation experiment and apply CONCAT to the results.
KeywordsInput Node Category Learning Connectionist Model Training Stage Error Frequency
Unable to display preview. Download preview PDF.
- 5.Lamberts, K. Feature-accumulation theory of speeded categorisation. Psychological Review in press.Google Scholar
- 6.Lamberts, K. & Freeman, R. P. J. Building object representations from parts: Tests of a stochastic sampling model. Journal of Experimental Psychology: Human Perception and Performance in press.Google Scholar
- 9.Peebles, D. The effect of stimulus frequency on classification accuracy and response time. Unpublished doctoral dissertation, University of Birmingham, Birmingham.Google Scholar