Artificial intelligence for mass appraisals of residential properties in Nicosia: Mathematical modelling and algorithmic implementation

Abstract

A recent study in property valuation literature, indicated that the vast majority of researchers and academics are focusing on Mass Appraisals rather than on further developing the existing methods. Researchers are using a variety of mathematical models from the field of Machine Learning and Artificial Neural Networks, which are applied to real estate valuations, with high accuracy. On the other hand, it appears that the professional valuers do no use those sophisticated models on their daily practice, using essentially the traditional 5 methods. At that point, authors deal with the ethical question that arises and that is whether those models can replace the judgment of the individual valuer. As in many other aspects of scientific research, and in particular in artificial intelligence applications, human intelligence is still dangerous to be replaced by machine intelligence (like i.e. the self-driving cars). Despite the fact that those models are proved to be extremely accurate in academic test cases, in real-world applications, they cannot be used without the audit of an experienced valuer. The aim of this work is to investigate the capabilities of such models and how they can be used in order to improve valuer's work.