Using a sample of 900 apartments from Cluj-Napoca, Romania, containing selling transactions for the second semester of 2019, and data for 33 locational, physical and neighbourhood-related attributes (socio-cultural, environmental, and urbanism related), our research objective is to test the performance in price prediction, and hence the utility, of the Artificial Neural Networking (ANN), as artificial intelligence model versus the Generalized Linear Model (GLM), as a regression model. By contributing to an ongoing debate, our empirical findings confirm the results of a predominant group of earlier studies, namely the superiority of ANN. Precisely, we found that ANN can better predict selling prices and provides stability of results. Additionally, we addressed the critiques related to the transparency of results, showing that ANN also has the ability to illustrate the significance of the different attributes of real estate, if appropriate statistical indicators are used. These findings can serve the different real estate valuation purposes, including that of the review of valuation reports.

Testing a model in property evaluation can be a difficult task due to the large variety of these models. The most popular models used in valuation are regression and neural networks. This paper applied a systematic review study and presents 11 types of regression models and 9 types of neural network models applied in real estate valuation. Our aim is to provide a tool for model selection applied in real estate valuation. The selection criteria were based on their applicability, user preferences and price estimation performance. The findings were slightly different from our expectations. Multi-Layer Perceptron (MLP) and Multiple Linear Regression (GLM) are the most applied and popular models in valuation.