SIMARK: AI SIMULATIONS FOR MARKETING

Abstract

The recent progress of artificial intelligence has boomed the development of explainable machine learning methods capable of extracting subtle insights from massive data sets and providing understandable explanations about them. The SIMARK project focuses on using: i) agent-based modeling (ABM) simulations; ii) explainable machine learning methods, and iii) advanced data visualization; in order to model and control complex marketing systems. These models comprise “virtual computational marketplaces” on a hitherto unprecedented scale to analyze, experiment with, feed with, and test against empirical real-world data.

SIMARK will overtake the current state of the art of this discipline. The scaling up of the ABM implementation to develop and analyze massive simulations will be faced. Likewise, explainable machine learning methods will be designed to uncover hidden problem-specific knowledge, allowing us to derive and validate generative theories from the simulation data explaining how complex behaviors emerge in markets. The automatic definition of marketing strategies to support the human decision making will also be tackled, considering the massive-scale, uncertain, and multicriteria problem nature. Advanced specific visualization techniques and methods to extract useful associations from these visualizations will be designed, permitting uncovering additional knowledge of the consumers’ behavioral patterns and the appropriate marketing campaign settings.

SIMARK becomes a truly interdisciplinary and innovative project that integrates disciplines such as complexity, ABM, explainable machine learning methods and human-centric modeling/decision making, big data, and information visualization for marketing. Although the current project is very challenging, the experience and expertise of the multidisciplinary research team and the collaboration with a SME of reference in the field, ROD Brand Consultants (ROD), ensures its successful development. SIMARK foresees the negotiation of exploitation agreements (including appropriate market prices) with the interested firms of the marketing sector.