Automating Cognitive Model Improvement by A{sup}* Search and Logistic Regression

摘要

A good cognitive model is important to the effectiveness to an intelligent tutor. In this paper we present a method of combining the A{sup}* search algorithm and logistic regression to automate the improvement of a cognitive model by 1) automatically generating different models by mutating learning factors in a base model 2) integrating logistical regression to evaluate different models 3)selecting the best model through a depth-first search algorithm.