Pemodelan Volatilitas Harga Komoditas Energi Dunia dengan Fractional Cointegration Model dan Support Vector Regression

Izati, Prajna Pramita (2023) Pemodelan Volatilitas Harga Komoditas Energi Dunia dengan Fractional Cointegration Model dan Support Vector Regression. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003212012-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (5MB) | Request a copy

Abstract

Komoditas energi merupakan hasil energi dalam bumi yang menghasilkan produk berupa sumber energi atau bahan bakar yang dapat diperdagangkan. Komoditas energi ini mencakup batu bara, gas alam, dan minyak mentah WTI serta minyak mentah Brent yang saling berkointegrasi satu sama lainnya. Komoditas energi menjadi input penting dalam industri dunia akibat itu tingginya ketergantungan dunia terhadap komoditas energi ini menjadikan perubahan harga komoditas ini menjadi penting untuk diperhatikan dan diramalkan. Penelitian ini bertujuan untuk membanding model GARCH/ARIMA, GARCH-Fractional Cointegration atau ARIMA-Fractional Cointegration, GARCH-SVR atau ARIMA-SVR untuk mendapatkan model yang paling sesuai untuk meramalkan volatilitas komoditas energi di mana kriteria pembandingnya yaitu sMAPE yang terkecil. Hasil uji heterogenitas pada data menunjukkan volatilitas return kuadrat dan realized volatility tidak homogen. Dalam analisis ini, residual dari kombinasi linear variabel komoditas energi mengikuti proses long memory yang stasioner menunjukkan hubungan cointegration antara variabel-variabel tersebut sehingga dapat dimodelkan dengan Fractional Cointegration dan Support Vector Regression. Berdasarkan hasil pemodelan didapatkan bahwa model GARCH adalah model terbaik untuk volatilitas return kuadrat komoditas batu bara, model GARCH-Fractional Cointegration adalah model terbaik untuk volatilitas return kuadrat minyak mentah Brent, dan model GARCH-SVR adalah model terbaik untuk minyak mentah WTI dan gas alam. Selain itu diperoleh juga bahwa model ARIMA merupakan model terbaik untuk realized volatility disjoint gas alam dan batu bara, model ARIMA-Fractional Cointegration merupakan model terbaik untuk realized volatility disjoint minyak mentah Brent, dan ARIMA-SVR adalah model terbaik untuk realized volatility minyak mentah WTI. Adapun untuk data realized volatility overlapping, model komoditas minyak mentah WTI, batu bara, dan gas alam diperoleh metode ARIMA merupakan metode terbaik. Sementara model minyak mentah Brent menggunakan metode ARIMA-SVR sebagai metode terbaik untuk meramalkan data realized volatility overlapping.
========================================================================================================================================
Energy commodities are products of energy in the earth that produce products in the form of tradable sources of energy or fuel. These energy commodities include coal, natural gas, and WTI crude oil as well as Brent crude oil which are cointegrated with one another. Energy commodities are important inputs in the world industry. As a result, the world's high dependence on these energy commodities makes changes in the prices of these commodities important to watch and predict. This study aims to compare the GARCH/ARIMA, GARCH-Fractional Cointegration or ARIMA-Fractional Cointegration, GARCH-SVR or ARIMA-SVR models to obtain the most suitable model for predicting the volatility of energy commodities where the comparison criterion is sMAPE which is the smallest. The results of the heterogeneity test on the data show that the return volatility is squared and realized volatility is not homogeneous. In this analysis, the residuals from the linear combination of energy commodity variables following a stationary long memory process show a cointegration relationship between these variables so that they can be modeled by Fractional Cointegration and Support Vector Regression. Based on the modeling results, it was found that the GARCH model is the best model for return squared volatility for coal, the GARCH-Fractional Cointegration model is the best model for return squared volatility for Brent crude oil, and the GARCH-SVR model is the best model for WTI crude oil and natural gas. In addition, it was also found that the ARIMA model is the best model for realized volatility disjoint natural gas and coal, the ARIMA-Fractional Cointegration model is the best model for realized volatility disjoint Brent crude oil, and ARIMA-SVR is the best model for realized volatility for WTI crude oil. As for realized volatility overlapping data, the commodity model for WTI crude oil, coal, and natural gas obtained by the ARIMA method is the best method. Meanwhile, the Brent crude oil model uses the ARIMA-SVR method as the best method for predicting realized volatility overlapping data.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Fractional Cointegration, GARCH, Komoditas energi, Pemodelan, Support Vector Regression, Energy commodities, Fractional Cointegration, GARCH, Modeling, Support Vector Regression
Subjects:	Q Science > QA Mathematics > QA274.2 Stochastic analysis Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) Q Science > QA Mathematics > QA280 Box-Jenkins forecasting Q Science > QA Mathematics > QA353.K47 Kernel functions (analysis) Q Science > QA Mathematics > QA402.3 Kalman filtering.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Prajna Pramita Izati
Date Deposited:	07 Aug 2023 07:48
Last Modified:	07 Aug 2023 07:48
URI:	http://repository.its.ac.id/id/eprint/104148

Actions (login required)

View Item