AI-Based on Machine Learning Methods for Urban Real Estate Prediction: A Systematic Survey

The advanced urban digitization enhancing a huge volume of data collected in many areas has led to the emergence of artificial intelligence (AI) based tools in decision support systems. These use various machine learning algorithms to extract valuable information for important decision-making such as house price predictions. The urban real estate investment business model is undergoing a fundamental overhaul attributed to digitization and a growing market for smart and environmentally demanding buildings. This technological breakthrough is reinforced by AI data analysis which greatly improves decision-making by anticipating price changes through predictive modelling. This issue is reinforced by the strong growth in the number of scientific papers published in recent years on this problem. Nevertheless, scarce effort has been made to assess what has been done thus far in order to identify the possibilities, the most popular or flexible techniques, the effect, and the challenges in order to expand the scope going forward. To fill this gap we evaluated 70 of the most relevant papers selected from the Scopus database. Overall, our study revealed a significant concentration of publications from the USA, China, India, Japan, and Hong Kong. These countries have the particularity not only to be very digitized and the more advanced research but also the higher stakes requiring the best decision-making. On the other hand, the data sizes used were often relatively small and the research areas of the authors would favour the strong use of simple ML methods over deep learning methods. Future research and applications should not only be enriched by the large and accurate data coming from the increased digitalization of cities and thus urban real estate but also address the explainability challenges of the models built. Addressing these non-exhaustive challenges would allow for better management of both research and business model design through a better understanding and use of intelligent decision support systems by real estate stakeholders.