Combining Psychological Models with Machine Learning to Better Predict People's Decisions.

摘要

Creating agents that proficiently interact with people is critical for many applications. Towards creating these agents, models are needed that effectively predict people's decisions in a variety of problems. To date, two approaches have been suggested to generally describe people's decision behavior. These models could either be based on theoretical rational behavior, or psychological models such as those based on bounded rationality. A second approach focuses on creating models based exclusively on observations of people's behavior. At the forefront of these type of methods are various machine learning algorithms. This paper explores how these two approaches can be compared and combined in different types of domains. In relatively simple domains, both psychological models and machine learning yield clear prediction models with nearly identical results. In more complex domains, psychological or machine learning alone cannot accurately predict people's decisions. However, improved models can be created by using machine learning techniques to refine parameters within psychological models. In the most complex domains, the exact action predicted by psychological models is not even clear, and machine learning models are even less accurate. Nonetheless, by creating hybrid methods that incorporate features from psychological models in conjunction with machine learning we can create significantly improved models for predicting people's decisions. To demonstrate these claims, we present a survey of previous and new results, taken from representative domains ranging from a relatively simple optimization problem, a more complex path selection domain, and complex domains of negotiation and coordination without communication.