NECSI Resources

 International Conference on Complex Systems (ICCS2007)

Agent-based Market Model Validation Using Empirical Data

Rosanna Garcia
Northeastern Univeristy

     Full text: Not available
     Last modified: June 29, 2007

Abstract
Agent-based marketing models (AMM) are becoming more commonly used in the marketing literature to represent individual actors (or groups) in a dynamic adaptive system (see forthcoming special issue on AMMs in the Journal of Business Research). Validating empirically-based AMMs is burdened with confounding from feedback loops, path dependencies, sensitivity to internal conditions and unpredictability of agent adaptation (Fagiolo, et al., 2005). Important questions being posed by agent-based modelers, especially those investigating ‘real-world’ marketing systems are: Which methods of validation are best? What levels should be considered? How does one know a model is correct?

In this paper we relate how empirical data may be used to validate AMMs focusing on diffusion using three different examples. One study used conjoint analysis to instantiate agents and to verify and validate the model based in the wine industry (Garcia, Rummel & Hauser, forthcoming), a second model uses survey data collected in conjunction with a US-based personal products manufacturer and the Bass diffusion model to verify and validate the model (Toubia, Garcia & Goldenberg, 2007) and the third uses survey data collected for a transportation industry related nested-logit model to accomplish validation (Garcia & Balbaky, 2007). We report on these different methods for using empirical data to conduct AMM verification and validation. This research contributes to the agent-based modeling simulation techniques that are more prevalent in the marketing literature.

References:

Fagiolo Giorgio, Moneta Alessio, Windrum Paul, Empirical validation of agent-based models: A critical survey. International Workshop on Agent-Based Modeling and Economic Policy Design University of Bielefeld, Germany, 2005.

Garcia, Rosanna, Paul Rummel & John Hauser, “Validating agent-based marketing models using conjoint analysis”, Journal of Business Research 2007.

Toubia, Olivier, Rosanna Garcia & Jacob Goldenberg, “Revisiting the Bass diffusion model using agent-based models”, Northeastern University working paper, 2007.

Garcia, Rosanna & Samer Balbaky, 2007, “The impact of green-house-gas emission policies on the diffusion of alternative fuel vehicles”, Northeastern University working paper, 2007.








Maintained by NECSI Webmaster Copyright © 2000-2007 New England Complex Systems Institute. All rights reserved.