Fuzzy inference system (FIS) is one of the artificial intelligence methodologies that works linguistically based on numerical data to reach effective categories as antecedent scientific results. This system considers also expert views in the forms of fuzzy logic sets. Pre-earthquake building evaluation studies have importance for each earthquake prone area to reduce the loss of property and human life. This paper provides the integration of probabilistic and FIS clustering procedures for the pre-earthquake assessment of each building. The essence of this methodology is to provide probabilistic risk and fuzzy c-means soft clustering methodological application. The fundamental variables are column area ratio, height ratio, frame effect, moment of inertia (stiffness) and building height. The proposed methodology is applied to more that thousand buildings in the Zeytinburnu region in Istanbul, Türkiye. Although visual and plan-based numerical data were obtained by examining more than a thousand buildings in this region, only fifty of them are presented for methodological applications. Five hazard classes are considered as ‘‘Collapse,” ‘‘Severe,” ‘‘Moderate’’, ‘‘Mild” and ‘‘Non.” For each variable at risk levels of 0.50, 0.20, 0.10, 0.04 and 0.01 corresponding hazard classes are obtained as “Collapse” 36%, “Severe” 26%, “Moderate” 28%, “Mild” 4% and “Non” 6%, respectively.
Zekâi Şen (2026) studied this question.