Perbandingan Kinerja ARIMAX dan Support Vector Regression (SVR) Dalam Prediksi Produk Domestik Bruto (PDB) di Indonesia

Marcelia, Firsta (2026) Perbandingan Kinerja ARIMAX dan Support Vector Regression (SVR) Dalam Prediksi Produk Domestik Bruto (PDB) di Indonesia. Undergraduate thesis, UPN Veteran Jawa Timur.

[img] Text (Cover)
21083010115.-cover.pdf

Download (2MB)
[img] Text (Bab 1)
21083010115.-bab 1.pdf

Download (268kB)
[img] Text (Bab 2)
21083010115.-bab 2.pdf
Restricted to Repository staff only until 12 November 2029.

Download (1MB)
[img] Text (Bab 3)
21083010115.-bab 3.pdf
Restricted to Repository staff only until 12 November 2029.

Download (727kB)
[img] Text (Bab 4)
21083010115.-bab 4.pdf
Restricted to Repository staff only until 12 November 2029.

Download (3MB)
[img] Text (Bab 5)
21083010115.-bab 5.pdf

Download (232kB)
[img] Text (Daftar pustaka)
21083010115.-daftarpustaka.pdf

Download (227kB)
[img] Text (Lampiran)
21083010115.-lampiran.pdf
Restricted to Repository staff only until 12 November 2029.

Download (242kB)

Abstract

Gross Domestic Product (GDP) is one of the primary indicators used to measure a country's economic growth. GDP fluctuations are influenced by various macroeconomic factors, such as inflation, interest rates, exports, and imports. GDP forecasting is important for governments, investors, and industry stakeholders in supporting more accurate and data-driven decision-making. Therefore, forecasting methods capable of capturing time series patterns and the effects of macroeconomic factors on GDP are required. This study aims to compare the performance of the Autoregressive Integrated Moving Average with Exogenous Variables (ARIMAX) and Support Vector Regression (SVR) methods in forecasting Indonesia's GDP. The data used in this study consist of quarterly data from 2012 to 2025 obtained from Statistics Indonesia (BPS), Bank Indonesia (BI), and the Ministry of Trade of the Republic of Indonesia. The variables analyzed include inflation, interest rates, exports, and imports as independent variables, while GDP serves as the dependent variable. In the SVR model, hyperparameter optimization was conducted using Grid Search and Bayesian Optimization, and the best-performing model was selected based on the evaluation results. Model performance was evaluated using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination (R²). The results show that the ARIMAX (4,1,3) model achieved an MAE of 199,785.08, an RMSE of 224,722.26, a MAPE of 3.41%, and an R² of 0.30. Meanwhile, the best SVR model obtained through Grid Search optimization with a linear kernel, C = 100, and epsilon = 0.1 achieved an MAE of 80,045.14, an RMSE of 99,970.14, a MAPE of 1.41%, and an R² of 0.86. Based on these evaluation results, the SVR method outperformed ARIMAX in forecasting Indonesia's GDP, as it produced lower prediction errors and a higher coefficient of determination.

Item Type: Thesis (Undergraduate)
Contributors:
ContributionContributorsNIDN/NIDKEmail
Thesis advisorHindrayani, Kartika MaulidaNIDN0009099205kartika.maulida.ds@upnjatim.ac.id
Thesis advisorTrimono, TrimonoNIDN0008099501trimono.stat@upnjatim.ac.id
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions: Faculty of Computer Science > Departemen of Data Science
Depositing User: Firsta Marcelia
Date Deposited: 13 Nov 2026 02:30
Last Modified: 13 Nov 2026 02:30
URI: https://repository.upnjatim.ac.id/id/eprint/61331

Actions (login required)

View Item View Item