Review of Housing Price Forecasting Methods Based on Machine Learning and Deep Learning

Abstract

Housing price forecasting holds profound significance in the development, investment, and policy formulation within the real estate market. Precise forecasts not only assist real estate developers in evaluating the potential value of projects but also furnish investors with a scientific basis to optimize their investment decisions. At the same time, home buyers can also gauge the price trend and appreciation potential of properties based on the forecast results. Moreover, governments and financial institutions rely on reliable housing price forecast results to formulate reasonable market supervision policies and control loan risks. This article discusses the status of house price forecasting in society in recent years and the latest research progress in applying machine learning and deep learning. Notably, it zeroes in on the role that the fusion of numerical features, textual descriptions, and image features plays in predicting housing prices. Through a methodical systematic of the existing research, this paper analyzes the characteristics, advantages and disadvantages of different models and their application scenarios. The results show that the fusion of multi-source data can significantly improve the accuracy of model prediction, but it also faces challenges in finding high-quality data and data processing and computing resources. Future studies should further optimize the feature extraction method to enhance the interpretability and adaptability of the model.