EVALUATION OF SEISMIC VULNERABILITY FOR NON-ENGINEERED BUILDINGS USING INTEGRATED ANALYTIC HIERARCHY PROCESS AND FUZZY INFERENCE SYSTEM

Abstract

Seismic vulnerability evaluation of non-engineered buildings (NEBs) in developing countries is hindered by data uncertainty, building diversity, and impractical conventional tools. This study proposes an integrated Analytic Hierarchy Process–Fuzzy Inference System (AHP–FIS) model to evaluate the seismic vulnerability of NEBs. Fourteen factors covering technical, socio-economic, and environmental dimensions were collected through field-based scoring and then processed using a Mamdani FIS model. A total of 120 observed post-earthquake house damage records were used to demonstrate the predictive accuracy of the model. The results show that AHP weighting indicated that the technical dimension (0.5151) was the dominant driver, followed by the environmental (0.2698) and socio-economic (0.2151) dimensions. The houses assessed fell to 47.5% of the High to Very High vulnerability classes. The model achieved 81.03% sensitivity, 83.87% specificity, 82.46% precision, and an overall accuracy of 82.50%, indicating that the model showed satisfactory classification performance and this is suitable as a preliminary regional screening and prioritization tool for seismic-risk mitigation of NEBs.