MODELING HOUSE SELLING PRICES IN JAKARTA AND SOUTH TANGERANG USING MACHINE LEARNING PREDICTION ANALYSIS

Abstrak

The increasing demand for housing in urban agglomerations, particularly in areas like Jakarta, has made homeownership a significant challenge for many, especially first-time buyers and the lower-middle class. Post-pandemic shifts have further influenced housing preferences, driving interest towards suburban areas with green spaces. Despite government efforts through mortgage subsidy programs, affordability remains a concern, particularly in peripheral regions. This study aims to analyze housing prices in various Jakarta regions using machine learning models, including Multiple Linear Regression (MLR), Support Vector Regression (SVR), Light Gradient Boosting Machine (LGBM), and Random Forest. A dataset of 554 house prices from West Jakarta, South Jakarta, Central Jakarta, and South Tangerang was used. The analysis focused on key predictors like land area, building area, bedrooms, and carports, with R² and Mean Squared Error (MSE) metrics evaluating model performance. Results showed that LGBM and Random Forest outperformed others with 0.8 R2 and low MSE, with building and land area as the most significant factors influencing prices. The study concludes that property size is a primary determinant of house prices, and there is a need for policy interventions to make housing more affordable. Additionally, apartment rentals offer a viable alternative, especially in central urban areas, where proximity to economic activities and facilities is crucial. The findings suggest that enhancing marketplace features with predictive tools could further assist buyers in making informed decisions.