Comparison of Regression Analysis with Machine Learning Supervised Predictive Model Techniques

Pardomuan Robinson Sihombing (BPS-Statistiscs Indonesia)
Sigit Budiantono (PKN-STAN, Indonesia)
Ade Marsinta Arsani (BPS-Statistics Indonesia)
Triana Mauliasih Aritonang (BPS-Statistics Indonesia)
Mohamad Arif Kurniawan (BPS-Statistics Indonesia)

Berdikari : Jurnal Ekonomi dan Statistik Indonesia
Vol 3 No 2 (2023): Berdikari: Jurnal Ekonomi dan Statistik Indonesia (JESI)

The happiness index is a parameter used to measure the level of happiness and well-being of people in a particular country or region. This research aims to determine the factors that contribute to people's happiness. In terms of modelling, this study will compare several regressions modelling using machine learning, including regression trees, random forests and Support Vector Regression (SVR). The SVR model has a minor error value in terms of MSE, RMSE and MAE compared to the other three models. The same thing happened when viewed from the value of R2 that the SVR model has an enormous value. This result indicates that SVR modelling is the best of the four models. A comprehensive policy is needed to increase a country's happiness index.