Function Learning: An Exemplar Account of Extrapolation Performance
Abstract
A function describes a one-to-one relationship between combinations of predictor and criterion variables. In this paper, we describe a new memory model that learns functional relationships. Two versions of the model are described. The first version learns the bivarite relationship between a single predictor and criterion. The second version expands on the first to multiple predictors. For both versions of the model, we present empirical data to test them and find that they do a good job of accounting for human performance.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 29, 2003
- Accession Number
- ADA593833
Entities
People
- Andrew Neal
- Peter J. Kwantes
Organizations
- Defence Research and Development Canada