Review and Outlook on Data-driven Real Estate Mass Appraisal

With the aim to effectively present the system and context of real estate mass appraisal, the article employs a literature induction method to systematically organize and summarize the research literature in this field. The findings indicate that: (1) In terms of mass appraisal models and improvements, new modern models, especially machine learning models, are widely applied and discussed; (2) In terms of model performance evaluation, the evaluation measures used in literatures include R², RMSE, MAPE, COD and the like; (3) In terms of data in mass appraisal, data scale and quality have become key issues in mass appraisal. By clarifying the current hotspots and trend, the article proposes that future research will focus on the integration of various models, the mining and use of multi-source data, and the practical application of mass appraisal models.