Influence of Social Overhead Capital Facilities on Housing Prices Using Machine Learning

Abstract

The South Korean residential real estate market is influenced by both the traditional dynamics of demand and supply and external factors such as housing policies and macroeconomic conditions. Considering the proportion of housing assets in individual wealth, market fluctuations can have significant implications. While previous studies have utilized variables such as GDP growth rate, patent issuance, and birth rate, and employed models such as LSTM and ARIMA for housing price predictions, many have overlooked the influence of local factors. In particular, there has been insufficient investigation into the impact of subway stations and living social overhead capital facilities on housing prices, especially in metropolitan areas. This study seeks to bridge this gap by analyzing the usage trends of subway stations, evaluating the impact of living social overhead capital facilities on housing values, and deriving the optimal machine learning model for price predictions near subway stations. We compared and analyzed a total of eight machine learning regression models, including Linear Regression, Decision Tree, Random Forest, LightGBM, Ridge, Lasso, Elastic Net, and XGBoost, all of which are popular regression models, especially in the context of machine learning and data science. Through comparative analysis of these machine learning techniques, we aim to provide insights for more rational housing price determinations, thereby promoting stability in the real estate market.

Keywords

• feature extraction
• housing price prediction
• living social overhead capital facilities
• subway proximity
• machine learning