Predicting Greater Jakarta Area House Prices Using Random Forest and Linear Regression

This study focuses on analyzing and predicting house prices in the Greater Jakarta Area using a machine learning approach, specifically comparing the performance of random forest regression and multiple linear regression. the increasing demand for adequate housing in Greater Jakarta Area, coupled with fluctuating house prices influenced by factors like land size, building size, number of bedrooms, bathrooms, and other facilities, necessitates an accurate price prediction system to assist both the public and businesses in decision-making. data was collected from Rumah123.com via Kaggle, followed by pre-processing and exploratory data analysis (EDA). the models were built using both algorithms and evaluated through 10-fold cross-validation, with an 80% training and 20% testing data split. the results demonstrate that random forest regression outperforms multiple linear regression, achieving a correlation coefficient of 0.5043 and a mean absolute error of 157,698,532. in contrast, multiple linear regression (m5p) yielded a correlation coefficient of 0.4895 and a mean absolute error of 209,890,933. therefore, random forest regression is recommended as a superior model for house price prediction in the Greater Jakarta Area region.