Discovering predictive variables when evolving cognitive models

Abstract

A non-dominated sorting genetic algorithm is used to evolve models of learning from different theories for multiple tasks. Correlation analysis is performed to identify parameters which affect performance on specific tasks; these are the predictive variables. Mutation is biased so that changes to parameter values tend to preserve values within the population's current range. Experimental results show that optimal models are evolved, and also that uncovering predictive variables is beneficial in improving the rate of convergence.