Estimasi Risiko Portofolio Optimal Pada Saham Idx-Mes Bumn 17 Berdasarkan Hasil Prediksi Harga Saham Menggunakan Bidirectional Long Short-Term Memory

Wulandari, Ratna Maulidah (2026) Estimasi Risiko Portofolio Optimal Pada Saham Idx-Mes Bumn 17 Berdasarkan Hasil Prediksi Harga Saham Menggunakan Bidirectional Long Short-Term Memory. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043221097-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Pesatnya perkembangan pasar saham di Indonesia mendorong meningkatnya partisipasi masyarakat, dengan jumlah investor bertambah lebih dari 57% antara akhir 2022 hingga Mei 2025. Peningkatan ini menunjukkan tingginya minat terhadap instrumen pasar modal sekaligus kesadaran akan pentingnya perencanaan keuangan jangka panjang. Sebagai negara dengan populasi Muslim terbesar di dunia, kebutuhan akan instrumen investasi sesuai prinsip syariah juga meningkat. Untuk menjawab kebutuhan tersebut, Bursa Efek Indonesia bersama Masyarakat Ekonomi Syariah meluncurkan indeks IDX-MES BUMN 17, yang terdiri dari saham BUMN dan afiliasinya dengan likuiditas tinggi, kapitalisasi besar, dan fundamental kuat. Penelitian ini bertujuan mengoptimalkan portofolio saham syariah IDX-MES BUMN 17 menggunakan model Mean Absolute Deviation (MAD) berdasarkan prediksi harga saham dari metode Bidirectional Long Short-Term Memory (Bi-LSTM), serta mengestimasi risiko portofolio melalui Value at Risk (VaR) dengan simulasi Monte Carlo. Data yang digunakan adalah harga penutupan harian 8 saham dengan kapitalisasi pasar terbesar dari indeks IDX-MES BUMN 17 selama periode 3 Maret 2023 hingga 30 September 2025. Hasil penelitian menunjukkan bahwa prediksi Bi-LSTM memiliki MAPE antara 1,101%–4,680%, menunjukkan akurasi sangat baik. Portofolio optimal Model Mean Absolute Deviation (MAD) tersusun atas 4 saham dengan bobot masing – masing saham yaitu PGAS 45,450%, BRIS 24,302%, ANTM 19,562%, dan PGEO 10,686%, dimana expected return yang dihasilkan lebih besar dari keseluruhan saham penyusunnya, dan risiko portofolio lebih kecil dari keseluruhan saham penyusunnya. Tingkat risiko portofolio MAD berdasarkan VaR yang disimulasikan menggunakan Monte Carlo dengan replikasi sebanyak 14.867.409 kali pada tingkat kepercayaan 95% adalah sebesar -2,2546%. Hasil penelitian ini dapat menjadi referensi bagi investor dalam pengambilan keputusan investasi yang optimal pada saham BUMN sesuai prinsip syariah.
============================================================================================================================
The rapid development of Indonesi a’s stock market has driven increased public participation, with the number of investors rising by more than 57% between the end of 2022 and May 2025. This growth reflects strong interest in capital market instruments as well as increased awareness of the importance of long-term financial planning. As the country with the largest Muslim population in the world, Indonesia has also experienced growing demand for investment instruments that comply with Sharia principles. To address this need, the Indonesia Stock Exchange, in collaboration with the Sharia Economic Society, launched the IDX-MES BUMN 17 index, which consists of BUMN and their affiliates with high liquidity, large market capitalization, and strong fundamentals. This study aims to optimize a Sharia-compliant stock portfolio from the IDX-MES BUMN 17 index using the Mean Absolute Deviation (MAD) model based on stock price predictions generated by the Bidirectional Long Short-Term Memory (Bi-LSTM) method, and to estimate portfolio risk using Value at Risk (VaR) through Monte Carlo simulation. The data used consist of daily closing prices of eight stocks with the largest market capitalization in the IDX-MES BUMN 17 index over the period from March 3, 2023, to September 30, 2025. The results show that the Bi-LSTM predictions achieve MAPE values ranging from 1.101% to 4.680%, indicating very high predictive accuracy. The optimal MAD portfolio consists of four stocks with weights of 45.450% for PGAS, 24.302% for BRIS, 19.562% for ANTM, and 10.686% for PGEO, producing an expected return higher than that of the individual constituent stocks while exhibiting lower portfolio risk. The MAD portfolio risk level based on VaR, simulated using Monte Carlo with 14,867,409 replications at a 95% confidence level, is −2.2546%. These findings can serve as a reference for investors in making optimal investment decisions in Sharia-compliant BUMN stocks.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Bi-LSTM, IDX-MES BUMN 17, MAD, Simulasi Monte Carlo, VaR
Subjects:	H Social Sciences > HA Statistics > HA31.7 Estimation
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Ratna Maulidah Wulandari
Date Deposited:	19 Feb 2026 01:14
Last Modified:	19 Feb 2026 01:14
URI:	http://repository.its.ac.id/id/eprint/132506

Actions (login required)

View Item