Impact of heuristic knowledge discovery techniques on multi-agent simulation of cultural evolution

The goal for of this research is to investigate and develop heuristic tools in order to extract meaningful knowledge or ontology from archeological large-scale data sets for the archaic state multi-agent based simulation of the Valley of Oaxaca. We are interested in developing rules that constrain the actions that the agents can take within their physical environment. The data used to constitute this ontology will be taken from the Oaxaca Valley Surface Survey. The knowledge is extracted the data set via Data Mining techniques.; This task is done via two evolutionary guided machine-learning techniques, Decision Tree Induction and Rough Sets. Decision Tree induction is guided by Cultural Algorithms while Rough Set Induction is guided by Genetic Algorithms. Decision Tree and Rough Set inductive learning techniques make different assumptions about the nature of the data and are designed to complement each other. The Decision Tree approach assumes that the values and precise whereas the Rough Set approach assumed some uncertainty in the data description. Each of these approaches will be used to extract rules and then thy are compared syntactically and semantically. The better rules will be used to constrain the agents' behavior within the environment.