Prediksi Harga Minyak Mentah Dunia Menggunakan Support Vector Regression Berbasis Genetic Algorithm, Particle Swarm Optimization, dan Grid Search

Suhartono, Putri Regina (2024) Prediksi Harga Minyak Mentah Dunia Menggunakan Support Vector Regression Berbasis Genetic Algorithm, Particle Swarm Optimization, dan Grid Search. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5006201009-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (2MB) | Request a copy

Abstract

Minyak bumi berperan besar dalam perekonomian global karena hasil olahannya sering digunakan dalam kehidupan sehari hari. Harga minyak dunia dapat naik turun tergantung tingginya permintaan global dan kebijakan yang dikeluarkan Organization of the Petroleum Exporting Countries (OPEC). Jika harga minyak naik maka dapat menimbulkan inflansi dan menurunkan daya beli konsumen. Sebaliknya, harga minyak yang turun dapat merugikan negara produsen minyak dan industri pengelola minyak. Oleh sebab itu, prediksi terhadap harga minyak mentah dunia menjadi penting sebagai upaya untuk mengelola volatilitas pasar dan dampaknya terhadap ekonomi global. Penelitian ini dilakukan untuk memprediksi harga minyak dunia mengunakan Support Vector Regression (SVR) dengan metode optimasi Genetic Algorithm (GA), Particle Swarm Optimization (PSO), dan Grid Search (GS) untuk mendapatkan parameter optimal. Data yang digunakan adalah OPEC Basket Price harian periode 2003 – 2023. Analisis SVR dilakukan menggunakan kernel linear, polynomial, dan Radial Basis Function (RBF), dimana kernel terbaik selanjutnya akan dioptimasi sehingga dihasilkan parameter cost (C), gamma (γ), dan epsilon (ε) yang optimal berdasar nilai RMSE yang terkecil. Hasil penelitian diperoleh bahwa kernel RBF merupakan yang terbaik dengan RMSE sebesar 2,120208 dengan nilai parameter C sebesar 1, γ sebesar 0,0001, dan ε sebesar 0,0001. Dari optimasi yang telah dilakukan diperoleh hasil bahwa metode SVR-GA memiki RMSE terkecil yaitu sebesar 2,119924 dengan nilai parameter C sebesar 0,9314789, γ sebesar 0,00008454813, dan ε sebesar 0,00006456384. Meski tidak terlalu signifikan, metode GA mampu mengoptimasi parameter SVR kernel RBF pada prediksi harga minyak mentah dunia.
=================================================================================================================================
Crude oil plays a significant role in the global economy because its processed products are frequently used in daily life. World oil prices can fluctuate depending on the high global demand and the policies issued by the Organization of the Petroleum Exporting Countries (OPEC). If oil prices rise, it can lead to inflation and reduce consumer purchasing power. Conversely, falling oil prices can disadvantage oil-producing countries and the oil processing industry. Therefore, predicting global crude oil prices becomes important as an effort to manage market volatility and its impact on the global economy. This research aims to predict world oil prices using Support Vector Regression (SVR) with optimization methods such as Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and Grid Search (GS) to obtain optimal parameters. The data used is the daily OPEC Basket Price for the period 2003 – 2023. SVR analysis is conducted using linear, polynomial, and Radial Basis Function (RBF) kernels, where the best kernel will be further optimized to produce optimal cost (C), gamma (γ), and epsilon (ε) parameters based on the smallest RMSE value. The research results show that the RBF kernel is the best with an RMSE of 2.120208, with parameter values of C at 1, γ at 0.0001, and ε at 0.0001. From the optimizations carried out, it was found that the SVR-GA method has the smallest RMSE, which is 2.119924, with parameter values of C at 0.9314789, γ at 0.00008454813, and ε at 0.00006456384. Although not very significant, the GA method can optimize the SVR RBF kernel parameters in predicting world crude oil prices.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Genetic Algorithm, Grid Search, Particle Swarm Optimization, Prediksi, Support Vector Regression
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA353.K47 Kernel functions (analysis)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User:	Putri Regina Suhartono
Date Deposited:	24 Jul 2024 01:37
Last Modified:	24 Jul 2024 01:37
URI:	http://repository.its.ac.id/id/eprint/108659

Actions (login required)

View Item