Providing a Forecasting Model in Assessing the Security of Urban Park Tourists Based on the Physical Structure of Urban Design

Security is one of the most important aspects of urban park design for tourists and citizens to achieve utmost satisfaction. The present study aims to provide a model for forecasting the security of urban parks based on the rate of crime occurrence in them; given the facilities, landscape quality and existing facilities in the parks using artificial neural networks that ultimately prioritize variables impacting the level of security in parks.

In   the implementation of the present study, a multilayer Perceptron network was used to process the data with the neural network intelligent tool. First of all, 99 urban parks in Tehran and Karaj were selected, and information on regional variables (2 variables) and physical-service (22 variables) in all parks were prepared through a field study. Regional variables including area of the park, the number of parks in the region and the population density of the region and service variables including the type of garden furniture, cultural and religious facilities and landscape quality variables were taken into consideration. The level of security of park visitors was determined based on the average number of reported violations per week, and modeling was done in MATLAB 2018. The value of model explanation coefficient (R2) in the whole data was 0.86 and in the network test data was 0.84, which indicates the exact application of artificial neural network in modeling the security of urban parks.

Results of the sensitivity analysis showed that the number of lighting bases, the number of sports levels, the number of water resources and the number of cultural and religious facilities, respectively, left the greatest impact on the safety of visitors in urban parks. Therefore, in the management of urban green spaces, paying attention to the design structure should be a priority to ensure environmental security.