Prediksi Indeks Ekonomi Hijau di Provinsi Jawa Timur Menggunakan Stacking Ensemble: XGBoost, LightGBM, dan CatBoost

Salsabilah, Andini Fitriyah (2025) Prediksi Indeks Ekonomi Hijau di Provinsi Jawa Timur Menggunakan Stacking Ensemble: XGBoost, LightGBM, dan CatBoost. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 21081010217.-cover.pdf Download (1MB)
	Text (bab1) 21081010217.-bab1.pdf Download (222kB)
	Text (bab2) 21081010217.-bab2.pdf Restricted to Repository staff only until 7 July 2027. Download (914kB)
	Text (bab3) 21081010217.-bab3.pdf Restricted to Repository staff only until 7 July 2027. Download (725kB)
	Text (bab4) 21081010217.-bab4.pdf Restricted to Repository staff only until 7 July 2027. Download (1MB)
	Text (bab5) 21081010217.-bab5.pdf Download (164kB)
	Text (daftarpustaka) 21081010217.-daftarpustaka.pdf Download (195kB)

Abstract

Indonesia is currently facing significant challenges in achieving sustainable development, particularly in balancing economic growth with environmental sustainability. Disparities among economic, social, and environmental aspects necessitate the presence of a comprehensive measurement tool to evaluate green economy performance. The Green Economy Index (GEI), introduced by the Ministry of National Development Planning (Bappenas), serves as a crucial instrument in assessing green economic achievements. However, data limitations at the provincial level, such as in East Java, hinder effective evaluation and policy formulation.This study proposes a machine learning-based approach to predicting GEI using a stacking ensemble algorithm that integrates three robust base models: XGBoost, LightGBM, and CatBoost. The model was developed using economic, social, and environmental indicators and evaluated using a holdout set to assess its accuracy and generalization capability.The results demonstrate that the stacking ensemble achieved the best performance with an RMSE of 0.0302, MAE of 0.0220, and R² of 0.9767, outperforming all individual models. For comparison, CatBoost recorded RMSE of 0.0317, MAE of 0.0229, and R² of 0.9743. LightGBM achieved RMSE of 0.0348, MAE of 0.0260, and R² of 0.9691. While XGBoost showed RMSE of 0.0413, MAE of 0.0315, and R² of 0.9564. These findings confirm that the stacking ensemble approach surpasses single models in accurately predicting the GEI. Thus, this approach can be a valuable data-driven decision support tool for advancing sustainable green economic development at the regional level.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Rahmat, Basuki	NIDN0023076907	basukirahmat.if@upnjatim.ac.id
Thesis advisor	Puspaningrum, Eva Yulia	NIDN0005078908	evapuspaningrum.if@upnjatim.ac.id

Subjects:

T Technology > T Technology (General)

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

Andini Fitriyah Salsabilah

Date Deposited:

07 Jul 2025 08:57

Last Modified:

07 Jul 2025 08:57

URI:

https://repository.upnjatim.ac.id/id/eprint/39266