"An Advanced Geospatial Analysis Model of Real Estate Assets Based on a Neural Network Approach

Real estate assets prices are a function of many parameters, physical and geographical, social and political, and others. These parameters, which usually have similar values within the same neighborhoods in urban areas, may be defined as general parameters. In spite of that fact ‐ of having similar parameters' values within the same neighborhoods ‐ we are facing different levels of real estate assets' prices in the same geographic areas. In order to analyze these general parameters and their effect on real estate assets' prices, the neural network model, combined with geographic information systems, is suggested in this paper as a tool to enable the estimating of assets' prices. As a geospatial tool, we limit ourselves in developing this model only to geographical and social parameters (henceforth, general parameters). As there are other data mining models to evaluate geospatial phenomena, the results of the neural network model are compared to a standard interpolation method, showing the advantages of the suggested method in estimating assets' real estate prices. In addition, as not all people have the same considerations in evaluating properties, the suggested model's implementation is extended for individual parameters, reflecting the preferences of people, to enable each person, through his own preferences, to define levels of real estate assets suited to his desires and abilities. Accordingly, the paper starts with a review of the general geographical parameters and their effect on real estate assets prices by presenting the mathematical model for evaluation and formalizing the prices and the weights of each parameter, using the neural network implementation with hidden layers, to formalize nonlinear models for calculating the parameters' weights and indicating real estate assets prices for the general geographical parameters. Then, individual parameters were taken into account to indicate the dynamic preferences regarding the real estate assets prices per person. The implementation, when given individual parameters from a specific person's viewpoint with their weights, in addition to the general parameters effect, enables the user to achieve satisfying answers for his own requirements.