Aesthetic quality modeling of the form of natural elements in the environment of urban parks

Assessing the aesthetic quality of the landscape, especially in natural ecosystems of urban areas, with the aim of utilization management, can be an effective step towards optimal and sustainable use of the urban parks' environment and services. The lack of studies on the form of landscape elements and combination of natural attributes in landscape beauty is a main shortcoming in this field. Therefore, the purpose of this study is to develop an artificial neural network model and tool to predict the aesthetic quality of the landscape containing water and plants in different forms and volume. Therefore, we recorded 6 landscape characteristics in 100 landscapes in 10 urban parks. Multi-layer perceptron model (MLP) represented the highest value of R-2 in training (0.98) and test (0.94) data set. According to the sensitivity analysis, the values of trees planting form, water, hedge, grass, water body form and trees were determined respectively as the inputs influencing the MLP model outputs. The urban park landscapes with more trees in the form of strip or group and water bodies and less hedge and grass would likely attract park visitors' attraction. The designed graphical user interface in MATLAB software, makes MLP an environmental decision support system tool for landscape designers and it is a platform to predict the quality of environment and the future researches should be on discussing or the relationship between spatial for of landscape attributes. In practice, the designed environmental decision support system tool is applied by landscape managers to predict the aesthetic quality of landscape in designing new parks.