Comparisons of Validated Agent-Based Model and Calibrated Statistical Model

Abstract

Agent-based models are generative models to provide the unrecorded, yet important storylines to model users. Such generative models might have less accuracy in the prediction, but the provided insights would be invaluable even compared to the accurate predictions. In spite of this different value proposition, outsiders as well as insiders of the agent-based modeling community often concern about the accuracy of these generative models. Hence, this chapter intends to investigate the differences of the model for regeneration of the real world, i.e., agent-based model, and the model for accurate predictions of the real world, i.e., statistical model. The investigations include qualitative as well as quantitative comparisons of the two types of models. Qualitatively, the two types of models are contrasted by what they can contribute at different levels, i.e., conceptual contribution, design contribution, and prediction contribution. Quantitatively, an agent-based model for city commerce was compared to the corresponding statistical model. This particular comparison, again, casts light on the trade-off of different contributions from different models.