CEREBRAL WEEVIL: A Machine Learning Model for Hemispheric Categorization of Complex Visual Patterns
Abstract
This report establishes the role of machine learning as a new human performance modeling technique. Learning mechanisms are modeled based on an empirical analysis of psychological reality and are subsequently tested empirically for their demonstration of that reality. Any variations in differences are considered in terms of changes to both future experimental studies of humans, and new evolvements in the programs used to model the reality. The reality used for this study involved hemispheric processing in learning visual patterns. The factors which influence hemispheric processing are modelled using the ID3 inductive machine learning technique. Various subsets of training data were used to assess the point of convergence between the research data and the model. The model successfully simulates the experimental data when it approaches a subset containing 90% of the examples. Conclusions look at ways to improve the model's convergence at a much lower percentage of examples induced. Keywords: Machine learning; Cognitive science; Cerebral cortex; Hemispheric asymmetry; Human performance modeling.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1989
- Accession Number
- ADA218143
Entities
People
- Michael D. Mcneese
Organizations
- Armstrong Laboratory