Explorations in Human and Machine Learning of Decision Trees

Abstract

We explore the boundaries of learnability, ecological rationality, and decision robustness in uncertain, non-stationary, finite-sample environments. Our approach combines machine-learning-based heuristic search techniques with the Integrated Learning Model (ILM) computational cognitive process theory of human and animal learning. The scientific contributions of this research are in understanding whether and how decision heuristics are acquired in binary classification contexts, with an emphasis on fast and frugal decision trees. The real world relevance of this research is in improved decision training and aiding.


Back to Table of Contents