06-Prediksi Harga Rumah Menggunakan Linear Regression pada Perumahan Kecamatan Paguyangan-Matien

Abstract

House prices in Indonesia fluctuate in line with economic growth. In Paguyangan District, factors such as land area, building area, number of rooms, electrical capacity, and proximity to main roads are key determinants of house prices. This study aims to assess the accuracy of a Linear Regression model in predicting house prices using 341 data points obtained through observation and interviews. The research stages include data pre-processing (data cleaning, categorical feature transformation, and outlier removal using the Z-Score method), data splitting using the 10-Fold Cross Validation technique, model training, and evaluation using R², MSE, RMSE, and MAE metrics. The best result was achieved in the sixth fold with an R² of 0.9427, RMSE of IDR 13,320,382.01, and MAE of IDR 10,541,293.61, indicating a relatively small prediction error compared to actual prices. This value indicates that 94.27% of the variation in house prices can be explained by the developed regression model. These findings conclude that the Linear Regression model in this study demonstrates excellent performance and is suitable for application in the context of house price prediction.