A clustering-based semi-automated technique to build cultural ontologies

Abstract

This article presents and validates a clustering-based method for creating cultural ontologies for community-oriented information systems. The introduced semiautomated approach merges distributed annotation techniques, or subjective assessments of similarities between cultural categories, with established clustering methods to produce “cognate” ontologies. This approach is validated against a locally authentic ethnographic method, involving direct work with communities for the design of “fluid” ontologies. The evaluation is conducted with of a set of Native American communities located in San Diego County (CA, US). The principal aim of this research is to discover whether distributing the annotation process among isolated respondents would enable ontology hierarchies to be created that are similar to those that are crafted according to collaborative ethnographic processes, found to be effective in generating continuous usage across several studies. Our findings suggest that the proposed semiautomated solution best optimizes among issues of interoperability and scalability, deemphasized in the fluid ontology approach, and sustainable usage.