<![CDATA[Penelitian ini bertujuan untuk menguji dan membandingkan efektifitas berbagai model machine learning dalam memprediksi harga rumah di wilayah Jabodetabek, yang merupakan kawasan dinamis dan berkembang pesat di Indonesia. Data dikumpulkan dari marketplace properti di Indonesia dan dilengkapi dengan indikator sosial-ekonomi dari Badan Pusat Statistik (BPS). Penelitian ini menguji lima model yaitu Ridge Regression, Lasso Regression, Elastic Net Regression, Artificial Neural Network (ANN), dan Random Forest (RF) dengan fokus pada kinerja masing-masing di berbagai transformasi data. Di antara kelima model tersebut, Random Forest menunjukkan kinerja paling unggul dengan nilai R² sebesar 0,8715 pada skala log-transformed dan 0,8242 pada skala asli, yang masing-masing menjelaskan sekitar 87% dan 82% variasi harga perumahan. Faktor-faktor penentu utama, seperti luas bangunan, luas tanah, dan lokasi, diidentifikasi sebagai variabel yang paling berpengaruh. Hasil penelitian ini memberikan wawasan berharga bagi pengembang properti, investor, dan pembuat kebijakan untuk menyusun strategi dalam pasar perumahan. Abstract This study aims to evaluate and compare the effectiveness of various machine learning models in predicting housing prices in the Jabodetabek region, a dynamic and rapidly developing area in Indonesia. Data were collected from Indonesia property marketplace and supplemented with socio-economic indicators from the Central Bureau of Statistics (BPS). The research examines five models—Ridge Regression, Lasso Regression, Elastic Net Regression, Artificial Neural Network (ANN), and Random Forest (RF)—with particular attention to model performance under different data transformations. Among these, the Random Forest model demonstrated superior performance, achieving an R² of 0.8715 on the log-transformed scale and 0.8242 on the original scale, thereby explaining approximately 87% and 82% of the variance in housing prices, respectively. Key determinants of housing prices, such as building area, land area, and location, were identified as the most influential factors. The findings offer valuable insights for property developers, investors, and policymakers to formulate more informed strategies in the housing market.]]>
