Peramalan Saham Perbankan Blue Chip Menggunakan CNN-LSTM dan Analisis Risiko dengan Simulasi Monte Carlo

Putri, Berlia Ferdhyta (2026) Peramalan Saham Perbankan Blue Chip Menggunakan CNN-LSTM dan Analisis Risiko dengan Simulasi Monte Carlo. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5006221073-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (6MB)

Abstract

Pasar modal Indonesia mengalami pertumbuhan dan semakin diminati masyarakat, tercermin dari meningkatnya jumlah investor saham mejadi 7.0001.268 SID pada Mei 2025. Saham merupakan instrument investasi yang menawarkan keuntungan tinggi, namun juga memiliki risiko yang signifikan akibat fluktuasi pasar. Pada sektor perbankan, empat saham blue-chip yaitu BBCA, BBRI, BMRI, dan BBNI tercatat memiliki kapitalisasi pasar besar masing-masing sebesar sekitar Rp1.229 triliun, Rp907 triliun, Rp669 triliun, dan Rp217 triliun pada Maret 2024. Fluktuasi harga saham yang tinggi membuat peramalan dan pengukuran risiko penting dilakukan, terutama bagi investor. Penelitian ini memprediksi harga saham menggunakan model Convolutional Neural Networks-Long Short-Term Memory (CNN-LSTM) dan mengukur risiko return saham menggunakan Value at Risk (VaR) berbasis 10.000 simulasi Monte Carlo. Data yang digunakan berupa harga penutupan harian periode 1 Juli 2020 hingga 31 Juli 2025. Hasil penelitian menunjukkan bahwa model CNN-LSTM memiliki performa prediksi yang sangat baik dengan nilai MAPE model terbaik masing-masing saham yaitu BBCA sebesar 1,684%, BBRI 2,163%, BMRI 2,462%, dan BBNI 2,468%. Estimasi VaR menunjukkan bahwa BBCA memiliki risiko terendah, BBRI dan BMRI berada pada kategori risiko menengah, sedangkan BBNI memiliki risiko tertinggi pada tingkat kepercayaan 90%, 95%, dan 99%. Seluruh perhitungan VaR dinyatakan akurat berdasarkan uji Kupiec dengan nilai LR < 3,814 dan p-value > 0,05, sehingga dapat digunakan investor sebagai dasar dalam membuat keputusan investasi. Penelitian ini memberikan insight bahwa saham BBRI cocok untuk investor konservatif, BBCA cocok untuk investor konservatif-moderat, BMRI untuk investor moderat, dan BBNI untuk investor agresif.
==================================================================================================================================
Indonesia’s capital market has demonstrated substantial growth and increasing public participation, as evidenced by the rise in stock investors reaching 7.001.268 SID as of May 2025. Stocks represent investment instruments that offer considerable return potential while simultaneously carrying significant risks due to inherent market volatility. Within the banking sector, the four blue-chip BBCA, BBRI, BMRI, and BBNI, recorded substantial market capitalizations of approximately IDR 1.229 trillion, IDR 907 trillion, IDR 669 trillion, and IDR 217 trillion, respectively, in March 2024. The pronounced fluctuations in stock prices underscore the need for rigorous forecasting and risk measurement, particularly for informed investment decision-making. This study employs a Convolutional Neural Networks-Long Short-Term Memory (CNN-LSTM) model to forecast stock prices and used Value at Risk estimated through 10.000 Monte Carlo simulations to measure stock return risk. The dataset consists of daily closing prices spanning the period from July 1st, 2020 to July 31st, 2025. The empirical findings indicate that the CNN-LSTM model achieves a high level of redictive accuracy, with MAPE values ranging from 1,68% to 2,85%. VaR estimation further reveals that BBCA is the lowest level of risk, BBRI and BMRI fall within the medium-risk category, whereas BBNI is the highest risk across the 90%, 95%, and 99% confidence intervals. All VaR estimates are validated through the Kupiec test, yielding LR values < 3,814 and p-values > 0,005, thereby confirming their statistical accuracy and suitability for investment risk assessment. Overall, this study provides strategic insights for investor, which BBRI is most appropriate for conservative risk profiles, BBCA for conservative-moderate profile, BMRI for moderate profiles, and BBNI for aggressive investors.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Convolutional Neural Network, Long Short-Term Memory, Peramalan Saham, Simulasi Monte Carlo, Value at Risk. Convolutional Neural Network, Long Short-Term Memory, Monte Carlo Simulation, Stock Price Forecasting, Value at Risk
Subjects:	H Social Sciences > HA Statistics > HA30.3 Time-series analysis H Social Sciences > HB Economic Theory > Economic forecasting--Mathematical models. Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) Q Science > QA Mathematics > QA76.9 Computer algorithms. Virtual Reality. Computer simulation.
Divisions:	Faculty of Mathematics, Computation, and Data Science > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User:	Berlia Ferdhyta Putri
Date Deposited:	15 Jan 2026 04:38
Last Modified:	15 Jan 2026 04:38
URI:	http://repository.its.ac.id/id/eprint/129637

Actions (login required)

View Item