Links of aesthetic value of multi-curvature artifacts, with their structural behavior, utilizing machine learning algorithms

Abstract

In this work, quantitative evidence on whether the aesthetic value of architectural design is associated with its structural response is investigated. The procedure for this study was based on 170 parametric projects, as well as their structural analysis results. The aesthetic value of an architectural project is an ambiguous concept, as it is based on each individual's perception of beauty. This is highlighted in the corresponding literature on parametric design, form-finding, and aesthetics, as well as in an extensive bibliometric analysis of several hundreds of research papers with respect to aesthetics and structures. Henceforth, a questionnaire was developed and given for vote to 23 persons from a variety of disciplines such as architects, engineers, as well as people with profession irrelevant to structures and/or aesthetics. The mean value of the votes of the questionnaire was considered as the dependent variable of the mathematical model, while the structural response parameters, internal forces and nodal displacements, were set as the Independent variables of the model. The obtained data were analyzed with regression analysis, utilizing linear and machine learning models, revealing a significant link between structural response and aesthetics.