Interactive estimation of agent-based financial markets models

Abstract

Building upon the interactive inversion method introduced by Ashburn and Bonabeau (2004), we show how to dramatically improve the results by exploiting modularity and by letting the computer learn user preferences.