A Statistical Assessment of Sinan’s Central Domed Mosques

This paper uses a statistical approach to assess the proportional relations between 47 central-domed mosque plans of the architect Sinan. The statistical method is a feature selection method that demonstrates both the significant as well as the redundant features within a dataset by conducting correlation and lasso regression analyses. The analysis dataset contains 16 building elements present in each central domed mosque plan, selected according to Sinan’s mosque ontology. The correlation analysis demonstrates 72 significant correlations; the lasso regression shows seven essential features in relation to the area of the central space. The study contributes to the field with an evaluation of both Sinan’s high- and low-level design decisions by presenting a statistical assessment methodology of a historical architectural dataset. Results show that Sinan sees the design of a mosque as both a structural and architectural design problem.