Statistical exploratory analysis of agent-based simulations in a social context

Nathalie Villa-Vialaneix (1, 2) , Christophe Sibertin-Blanc (3) , Pascal Roggero (4)

1 SAMM - Statistique, Analyse et Modélisation Multidisciplinaire (SAmos-Marin Mersenne)
2 UPVD - IUT - Département STID - Carcassonne
3 IRIT-SMAC - Systèmes Multi-Agents Coopératifs
4 LEREPS - Laboratoire d'Etude et de Recherche sur l'Economie, les Politiques et les Systèmes Sociaux

This paper presents a case study for motivating the use of data mining in a social agent-based simulation framework. The data used in this article come from the social simulation platform SocLab. They were generated from a model designed to analyze a real situation related to the management of a river in South West of France. Several standard statistics methods are used to analyze the possible outcomes of the discussion between the actors. Finally, a typology of these outcomes is obtained with self-organizing map.

Mots clés: agent-based modeling; self-organizing maps; organizational behaviour; SocLab

Mots clés: typology; social simulation

Domaines: Système multi-agents [cs.MA]; Applications [stat.AP]; Sociologie

Nathalie Villa-Vialaneix, Christophe Sibertin-Blanc, Pascal Roggero. Statistical exploratory analysis of agent-based simulations in a social context. Case Studies in Business, Industry and Government Statistics, 2014, 5 (2), pp.132-149. ⟨hal-02629776v2⟩