Application of Data Mining Techniques to Rural Vernacular Buildings: A Methodology for Characterisation and Awareness

Abstract

ABSTRACTData-mining techniques (DMTs) have been widely used in the context of existing buildings. However, studies comparing its application to different objects and using rural territories with vernacular constructions as case studies remain highly unexplored. This paper discusses a methodology applied towards characterising two rural settlements in Portugal and Spain using DMTs and GIS-based maps to support the interpretation of results. Conclusions such as the existence of outliers or the optimal number of clusters are drawn, as well as the nature of clusters and their pertinence in identifying buildings closer to the vernacular matrix. Adapting the numerical database is suggested as future research, reducing the number of variables to obtain more accurate insights into the identification of vernacular buildings. Investigations such as the one presented in this paper promote awareness concerning the conservation and improvement of rural settlements, ultimately stimulating action towards territorial cohesion and sustainable development through future rehabilitation actions in vernacular built heritage.