UC EVENTS

Real Estate Appraisal based on Machine Learning and Computer Vision 

The real estate market in Australia is an agent’s market. The real estate agent does the most of work for their customers which including evaluating property values, listing properties on public websites, finding successful buyers and so on. This process has been used for many years not just in Australia, but also in other countries. However, recently with powerful artificial intelligence (AI) and Machine learning technologies, it is possible to develop new decision support and assistive technologies, for improving the real estate appraisal process in terms of work efficiencies for real estate agents, as well as the buyers' experience. In this research, a novel price prediction framework is proposed, based on  combining the aesthetic information in property images and the attribute features from quantitative data, and with deep learning based AI and computer vision algorithms, it was possible to improve the prediction performance as compared to previously proposed approaches. Some of the other contributions of this research include, development of a new ACT real estate database, a mobile app for data collection, and a peer reviewed conference paper accepted in an International conference on Computational Intelligence in Xiamen, China (http://ssci2019.org/).