ANALISIS RISIKO MONTE CARLO DALAM PREDIKSI HARGA SAHAM MENGGUNAKAN KOMBINASI FUZZY TIME SERIES DAN LONG SHORT-TERM MEMORY

Rizki, Alfi Hidayatur (2025) ANALISIS RISIKO MONTE CARLO DALAM PREDIKSI HARGA SAHAM MENGGUNAKAN KOMBINASI FUZZY TIME SERIES DAN LONG SHORT-TERM MEMORY. Undergraduate thesis, UPN "Veteran" Jawa Timur.

	Text 21083010001-cover.pdf Download (1MB)
	Text (BAB I) 21083010001-bab1.pdf - Accepted Version Download (282kB)
	Text (BAB II) 21083010001-bab2.pdf Restricted to Repository staff only Download (630kB)
	Text (BAB III) 21083010001-bab3.pdf Restricted to Repository staff only Download (684kB)
	Text (BAB IV) 21083010001-bab4.pdf Restricted to Repository staff only Download (2MB)
	Text 21083010001-bab5.pdf Download (228kB)
	Text 21083010001-daftarpustaka.pdf Download (240kB)
	Text 21083010001-lampiran.pdf Restricted to Repository staff only Download (241kB)

Abstract

The fluctuating and dynamic nature of stock prices presents significant challenges in forecasting and risk management processes. The limitations of conventional prediction methods in capturing non-linear patterns necessitate the development of more adaptive approaches. Therefore, this study proposes the integration of statistical methods with deep learning architectures. The objective of this research is to develop a stock price prediction model for PT Indofood Sukses Makmur Tbk using a hybrid approach combining Fuzzy Time Series and Long Short-Term Memory (LSTM), as well as to implement risk analysis through Monte Carlo simulation in the measurement of Value at Risk (VaR). Historical stock price data were obtained from Yahoo Finance using web scraping techniques and underwent preprocessing, including the handling of missing values, duplicates, outliers, and feature selection. The evaluation results indicate that the combination of Fuzzy Time Series and LSTM yields highly accurate predictions, as evidenced by a Mean Absolute Percentage Error (MAPE) of 0.037%. Risk analysis was conducted using Monte Carlo simulation on both historical and predicted log return data. For historical data, the Cornish-Fisher approach was applied, resulting in an estimated VaR of Rp53,199.21, while for predicted data, the z-alpha approach produced a VaR of Rp40,068.10. The simulation results were validated using the Kupiec test. This study also developed a web-based prediction and risk analysis system with an intuitive user interface, built using the Flask framework. Results from black-box testing confirmed that all core system functions operate correctly.

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Idhom, Mohammad	NIDN0010038305	idhom@upnjatim.ac.id
Thesis advisor	Saputra, Wahyu Syaifullah Jauharis	NIDN0725088601	wahyu.s.j.saputra@if.upnjatim.ac.

Subjects:

Q Science > QA Mathematics > QA76.6 Computer Programming

Divisions:

Faculty of Computer Science > Departemen of Data Science

Depositing User:

Alfi Hidayatur Rizki

Date Deposited:

19 Sep 2025 03:52

Last Modified:

19 Sep 2025 03:52

URI:

https://repository.upnjatim.ac.id/id/eprint/43800