Recent theories on urban economy have highlighted how investments in transport system can improve
accessibility to certain locations and affect property values. Moreover, city users are placing greater attention on the
quality of the living environment. In polluted cities the environment quality has become an important attribute
Artificial Neural Network models and the data set
Artificial Neural networks can be used to predict the house sale price. This work highlights the role of the
environment quality considering the real estate price, accessibility and other local land-use variables. The proposed
model has been calibrated using data collected in the city of Taranto. Furthermore, a sensitivity analysis has been
carried out in order to identify the most significant input variables. The traditional multiple regression models and
the estimated neural network models were useful in highlighting how different transport characteristics as well as
the environmental quality affect the prices of real estate properties. The proposed model has a good fit in both
training and test results. The results provided by the neural network can support investments in transport system.
The ANN model can help appraisers in making assessments and environmental regeneration. However, further
investigations are ongoing. In particular, clustering methods will be applied to improve the statistical performance of
the ANN in order to capture specific characteristic of groups of properties.