Exploring Deep Learning and Regression Models for Real Estate Price Prediction: A Survey of Current Approaches

Abstract

Accurately predicting real estate prices is crucial for market stakeholders, including investors, policymakers, and buyers, as it aids in decision-making and risk management. Recent advancements in deep learning and regression models have significantly enhanced the ability to analyze the complex and multifaceted factors influencing real estate prices. This survey provides a comprehensive review of current approaches, focusing on deep learning techniques, such as convolutional neural networks (CNNs), recurrent neural networks (RNNs), and hybrid architectures, as well as traditional and advanced regression models, including linear regression, random forest regression, and gradient boosting methods. The study evaluates the strengths, limitations, and performance metrics of these models in handling diverse datasets, including structured data (e.g., property attributes, economic indicators) and unstructured data (e.g., images, text from listings). Additionally, the survey examines key challenges, such as data quality, feature engineering, and model interpretability, while highlighting emerging trends, such as automated feature selection and explainable AI. By synthesizing insights from recent research, this paper offers a roadmap for future studies to address existing gaps and improve the predictive accuracy and robustness of real estate price prediction models.