Peramalan Volatilitas Return IHSG Menggunakan GARCH-Genetic Algorithm-Support Vector Regression

Muliabanta, Natanael Hadi (2025) Peramalan Volatilitas Return IHSG Menggunakan GARCH-Genetic Algorithm-Support Vector Regression. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5006201107-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2027. Download (4MB) | Request a copy

Abstract

Volatilitas merupakan salah satu konsep penting dalam ekonometrika yang merujuk pada tingkat perubahan harga aset dalam periode waktu tertentu. Dalam konteks investasi, pemahaman terhadap volatilitas sangat krusial karena mempengaruhi keputusan investasi dan strategi pengelolaan risiko. Indeks Harga Saham Gabungan (IHSG) sering digunakan sebagai indikator kondisi pasar saham di Indonesia, di mana fluktuasi indeks ini dapat memengaruhi perilaku investor. Model GARCH yang diperkenalkan oleh Bollerslev merupakan pengembangan dari Model ARCH dari Engle, dimana model GARCH memiliki kelebihan dalam menangkap volatilitas yang berkepanjangan, namun akurasi prediksi dari GARCH yang kurang baik mendorong orang untuk mencari metode peramalan volatilitas yang lebih baik. Dalam upaya mendapatkan prediksi volatilitas terbaik model GARCH dikombinasikan dengan model regresi dari Support Vector Machine. Permodelan Support Vector Regression (SVR) dilakukan menggunakan ketiga kernel linear, polynomial, dan RBF untuk memprediksi nilai koreksi dari prediksi GARCH. Kernel dengan hasil prediksi terbaik akan dioptimasi menggunakan Genetic Algorithm (GA) untuk mendapatkan hyperparameter yang optimal, sehingga didapatkan prediksi hybrid dari model GARCH-GA-SVR. Hasil prediksi hybrid terbaik pada penelitian ini jatuh pada model GARCH-SVR dengan kernel RBF yang memberikan peningkatan pada hasil evaluasi MSE 54% lebih kecil daripada model GARCH. Model dengan kernel RBF kemudian dioptimasi menggunakan GA, sehingga didapatkan hyperparameter optimalnya C = 0,7454181 , Ɛ = 0,4777241; y = 0,04280406 . Model GARCH-GA-SVR memberikan peningkatan hasil evaluasi MSE 56% lebih kecil dari model GARCH dan 4,4% lebih kecil dari model terbaiknya GARCH-SVR, sehingga dapat disimpulkan Model GARCH-GA-SVR merupakan model yang lebih baik dalam memprediksi fluktuasi pada IHSG.
=================================================================================================================================
Volatility is an important concept in econometrics that refers to the rate of change in asset prices over a specific period. In the context of investment, understanding volatility is crucial as it influences investment decisions and risk management strategies. The Indonesian Composite Index (IDXComposite) is often used as an indicator of the stock market condition in Indonesia, where fluctuations in this index can affect investor behavior. The GARCH model, introduced by Bollerslev, is an extension of Engle's ARCH model, and while GARCH has the advantage of capturing persistent volatility, its prediction accuracy is sometimes inadequate, prompting researchers to seek better volatility forecasting methods. In an effort to obtain the best volatility prediction, the GARCH model is combined with the regression model of Support Vector Machine. Support Vector Regression (SVR) is performed using three kernels: linear, polynomial, and RBF to predict the correction of GARCH predictions. The kernel with the best prediction results is then optimized using Genetic Algorithm (GA) to obtain the optimal hyperparameters, leading to the hybrid prediction from the GARCH-GA-SVR model. The best hybrid prediction result in this study was achieved by the GARCH-SVR model with the RBF kernel, which showed a 54% improvement in MSE evaluation compared to the GARCH model. The RBF kernel model was then optimized using GA, obtaining it’s optimal hyperparameters C = 0,7454181 , Ɛ = 0,4777241; y = 0,04280406. The GARCH-GA-SVR model provided a 56% improvement in MSE evaluation compared to the GARCH model and a 4.4% improvement compared to the best GARCH-SVR model, concluding that the GARCH-GASVR is the superior model in predicting the volatility of the IDXComposite.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	GARCH, Genetic Algorithm, Hybrid, SVR, Volatilitas, Volatility.
Subjects:	Q Science > Q Science (General) > Q180.55.M38 Mathematical models Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression Q Science > QA Mathematics > QA353.K47 Kernel functions (analysis) Q Science > QA Mathematics > QA402.5 Genetic algorithms. Interior-point methods.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User:	Natanael Hadi Muliabanta
Date Deposited:	15 Jan 2025 03:31
Last Modified:	15 Jan 2025 11:42
URI:	http://repository.its.ac.id/id/eprint/116296

Actions (login required)

View Item