Peramalan Harga Nikel Indonesia Menggunakan Metode ARIMA, Support Vector Regression (SVR), dan Genetic Algorithm-Support Vector Regression (GA-SVR)

Najwa, Nevisra (2024) Peramalan Harga Nikel Indonesia Menggunakan Metode ARIMA, Support Vector Regression (SVR), dan Genetic Algorithm-Support Vector Regression (GA-SVR). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003201124-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (2MB) | Request a copy

Abstract

Kemajuan teknologi mendorong terjadinya transisi energi yang ramah lingkungan, khususnya energi listrik. Nikel merupakan salah satu komponen atau bahan utama yang sangat diperlukan dalam produksi baterai kendaraan listrik oleh dunia. Indonesia menjadi salah satu penyumbang nikel terbesar dan mempunyai peranan yang sangat strategis. Terdapat potensi yang tinggi bagi industri nikel Indonesia dan menjadikan sektor nikel sebagai peluang dalam berinvestasi. Harga nikel cenderung bervariasi dari tahun ke tahun dan menunjukkan pola fluktuatif setiap bulannya. Data harga nikel umumnya bersifat nonlinier, sehingga peramalan menggunakan ARIMA cenderung menghasilkan error yang cukup besar. Namun, model ARIMA sangat baik ketepatannya untuk peramalan jangka pendek. Penggunaan model SVR (Support Vector Regression) dalam peramalan memiliki performansi yang baik dalam mengatasi permasalahan time series dan kasus nonlinieritas pada data. Dalam penelitian ini dilakukan peramalan harga nikel di Indonesia menggunakan metode ARIMA dan Support Vector Regression (SVR). Penelitian ini menggunakan Genetic Algorithm (GA) untuk mengoptimasi hyperparameter SVR agar diperoleh model yang lebih baik dengan error yang lebih rendah. Ketiga metode akan dibandingkan menggunakan kriteria RMSE dan sMAPE menghasilkan peramalan yang lebih akurat. Berdasarkan hasil analisis yang telah dilakukan, model yang menghasilkan prediksi terbaik adalah SVR dengan optimasi algoritma genetika (GA-SVR) karena memiliki nilai RMSE in sample, RMSE out sample, dan SMAPE out sample yang paling kecil dibandingkan model lainnya.
============================================================
Technological advancements are driving the transition to environmentally friendly energy, particularly electric energy. Nickel is a crucial component needed for the production of electric vehicle batteries worldwide. Indonesia is one of the largest contributors of nickel and holds a very strategic role. There is high potential for the Indonesian nickel industry, making the nickel sector a promising investment opportunity. Nickel prices tend to vary from year to year and exhibit a fluctuating pattern monthly. Nickel price data is generally non-linear, making ARIMA forecasting prone to significant errors. However, the ARIMA model is highly accurate for short-term forecasting and for non-stationary time series data when linear. Using models like SVR (Support Vector Regression) in forecasting performs well in addressing time series problems and non-linearity in the data. This study forecasts nickel prices in Indonesia using ARIMA and Support Vector Regression (SVR) methods. The study employs a Genetic Algorithm (GA) to optimize SVR hyperparameters to achieve a better model with lower errors. The three methods will be compared using RMSE and sMAPE criteria to produce more accurate forecasts. Based on the analysis, the best-performing model is SVR with Genetic Algorithm optimization (GA-SVR), as it has the lowest in-sample RMSE, out-sample RMSE, and out-sample sMAPE compared to other models.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	ARIMA, Energy, Genetic Algorithm, Nickel, Support Vector Regression, ARIMA, Energi, Genetic Algorithm, Nikel, Support Vector Regression.
Subjects:	H Social Sciences > HA Statistics H Social Sciences > HA Statistics > HA31.3 Regression. Correlation Q Science Q Science > QA Mathematics Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) Q Science > QA Mathematics > QA402.5 Genetic algorithms. Interior-point methods.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Najwa Nevisra
Date Deposited:	09 Aug 2024 01:37
Last Modified:	09 Aug 2024 01:37
URI:	http://repository.its.ac.id/id/eprint/114934

Actions (login required)

View Item