Exploring safety determinants in historic urban areas of Iran: an artificial neural network study in Ardakan historic fabric

This research aims to identify the most important predictors of residents' satisfaction with urban safety in historical urban areas. Using advanced machine learning algorithms, a case study was conducted on the old city centre of Ardakan in Iran. The study analyzed literature to identify important variables related to urban safety and collected data were processed using Artificial Neural Networks (ANNs) to discover the patterns underlying them. Results revealed that physical and design-oriented issues such as mixed land use and diversity of activities, street permeability and readable pattern, and formal and informal surveillance were the most critical aspects affecting residents' satisfaction with urban safety. The study proposes some design and operational measures based on expert knowledge and outcomes of the study to improve urban safety in historical urban areas. Overall, the research suggests that urban safety is a multi-faceted concept that requires careful consideration of physical and design-oriented factors alongside social and environmental factors.