Fatony, Hafizh Ahsan (2025) Analisis Resiko Return Saham Internasional Sektor Teknologi menggunakan Value at Risk Peaks Over Threshold dan Conditional Value at Risk dengan Pendekatan Quantile Regression Neural Network. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5003211123-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5003211123-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy |
Abstract
Penelitian ini bertujuan untuk menganalisis dan membandingkan estimasi risiko return saham sektor teknologi internasional yang terdaftar di NASDAQ, yaitu Nvidia, Apple, Microsoft, dan Broadcom, selama periode 3 Januari 2022 hingga 3 Januari 2025. Metode yang digunakan mencakup Value at Risk (VaR) dengan pendekatan Extreme Value Theory Peaks Over Threshold (EVT-POT) berbasis moving window sepanjang 63 hari, serta Conditional Value at Risk (CoVaR) yang dihitung berdasarkan hasil estimasi VaR tersebut, namun menggunakan Regresi Kuantil Linier (QR) dan Regresi Kuantil Neural Network (QRNN) tanpa penerapan moving window secara langsung dalam modelnya. Hasil backtesting melalui Expected Shortfall dan Kupiec Test menunjukkan bahwa pendekatan CoVaR-QRNN memiliki performa terbaik dalam menangkap risiko ekstrem pada distribusi return saham, dengan nilai Expected Shortfall yang mendekati kuantil 5% dan validitas statistik yang tinggi. Sebaliknya, pendekatan VaR-EVT dianggap tidak valid karena hasil uji Kupiec menolak hipotesis nol, yang menyatakan bahwa frekuensi pelanggaran berbeda dengan tingkat risiko (τ). Selain itu, model CoVaR-QRNN memberikan estimasi yang lebih stabil dibandingkan QR linier. Hasil penelitian ini menunjukkan bahwa model nonlinier QRNN lebih adaptif terhadap dinamika pasar dan kompleksitas distribusi return saham, sehingga layak dijadikan sebagai alternatif unggulan dalam pengelolaan risiko portofolio dan pengambilan keputusan investasi.
======================================================================================================================================
This study aims to analyze and compare risk estimation of stock returns in the international technology sector listed on NASDAQ, namely Nvidia, Apple, Microsoft, and Broadcom, during the period from January 3, 2022 to January 3, 2025. The methods employed include Value at Risk (VaR) using the Extreme Value Theory Peaks Over Threshold (EVT-POT) approach based on a 63-day moving window, as well as Conditional Value at Risk (CoVaR), which is computed based on the estimated VaR but using Linear Quantile Regression (QR) and Quantile Regression Neural Network (QRNN) without applying a moving window directly in the model. Backtesting results using Expected Shortfall and the Kupiec Test show that the CoVaR-QRNN approach performs best in capturing extreme risks in the distribution of stock returns, with Expected Shortfall values that closely approximate the 5% quantile and demonstrate strong statistical validity. In contrast, the VaR-EVT approach is deemed invalid as the Kupiec test rejects the null hypothesis, indicating that the violation frequency differs from the target risk level (τ). Additionally, the CoVaR-QRNN model provides more stable estimates compared to linear QR. These findings suggest that the nonlinear QRNN model is more adaptive to market dynamics and the complexity of stock return distributions, making it a strong alternative for portfolio risk management and investment decision-making under global economic volatility.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Analisis Risiko Saham, Return, Value at Risk, Conditional Value at Risk, Regresi Kuantil Neural Network Stock Risk Analysis, Return, Value at Risk, Conditional Value at Risk, Quantile Regression Neural Network |
| Subjects: | Q Science Q Science > QA Mathematics Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) |
| Divisions: | Faculty of Mathematics and Science > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Hafizh Ahsan Fatony |
| Date Deposited: | 31 Jul 2025 10:27 |
| Last Modified: | 31 Jul 2025 10:27 |
| URI: | http://repository.its.ac.id/id/eprint/125214 |
Actions (login required)
 |
View Item |