Wang Xiaomin
ABSTRACT
With the growth of the Internet and the increasing dynamism of the real estate sector, particularly the second-hand condominium market, the ability to accurately forecast price trends has become increasingly important. However, price prediction remains challenging due to information asymmetry, fluctuating market conditions, and the complex interplay of economic, environmental, and property-specific factors. To address these concerns, this study designed and developed the Pre-Owned Condominium Unit Price Trend Forecasting System with Decision Support. The system integrates data acquisition, processing, analysis, and visualization functions into a single decision-support platform. Python was used for web data crawling and preprocessing, with cleaned datasets stored in a structured database. The backend was developed using the Flask framework to support modular and scalable system operations. Data mining techniques, including clustering and classification, were employed using algorithms such as K-Means and Decision Trees to identify key price-influencing variables and construct the predictive model. While the initial integration of XGBoost improved model performance, its prediction accuracy showed certain limitations when applied to highly heterogeneous datasets, indicating a need for further model optimization. The system offers multiple analytical features, including historical market visualization, comparative property evaluation, and price trend forecasting, enabling users to make more informed purchasing or investment decisions. Evaluation based on the ISO/IEC 25010 Standard revealed a high level of system quality in terms of functionality, reliability, usability, and maintainability. Recommendations for future enhancement include incorporating advanced ensemble learning algorithms, real-time data synchronization, and user-interface personalization to further improve decision support accuracy and user experience. Overall, the developed system contributes a practical and scalable solution for real estate market analysis, supporting stakeholders such as buyers, sellers, brokers, and analysts in navigating the complexities of the second-hand condominium market.
Keywords: Pre-Owned Condominium Unit, price trend forecasting system, decision support, XG boost method
https://doi.org/10.57180/bfhn7701