Peramalan Harga Saham Dari Portofolio Optimal Jakarta Islamic Index (Jii) Berbasis Metaheuristic Dan Gated Recurrent Unit (Gru)

Salsabila, Amanda Shofiyah (2025) Peramalan Harga Saham Dari Portofolio Optimal Jakarta Islamic Index (Jii) Berbasis Metaheuristic Dan Gated Recurrent Unit (Gru). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043221126_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (2MB) | Request a copy

Abstract

Pasar modal berperan penting dalam mendukung pertumbuhan ekonomi, termasuk di Indonesia. Preferensi masyarakat terhadap saham syariah semakin meningkat karena kinerjanya yang relatif stabil. Salah satu indeks syariah adalah Jakarta Islamic Index (JII) yang terdiri dari 30 saham syariah paling likuid. Namun, fluktuasi harga dan risiko investasi tidak dapat dihindari sehingga strategi diversifikasi melalui pembentukan portofolio optimal memiliki peran penting. Penelitian ini menggunakan metode Particle Swarm Optimization (PSO) sebagai pendekatan metaheuristic untuk optimasi portofolio dengan memaksimalkan Sharpe Ratio, serta model Gated Recurrent Unit (GRU) sebagai metode peramalan. Hasil penelitian menunjukkan bahwa portofolio optimal terbentuk melalui eliminasi berbasis return bebas risiko dan ESG Risk Rating sehingga hanya tersisa saham dengan prospek yang layak dan berkelanjutan. Karakteristik kandidat portofolio menunjukkan bahwa AKRA memiliki return stabil dengan volatilitas rendah, BRIS memiliki return tinggi namun lebih volatil, EXCL memiliki return sedang dengan risiko rendah, PGAS memiliki return baik dengan risiko sedang, serta PGEO memiliki return tinggi dengan risiko besar. Korelasi antar saham relatif rendah sehingga peluang diversifikasi tetap terbuka. Optimasi PSO menghasilkan empat saham dalam portofolio optimal yaitu PGAS, BRIS, PGEO, dan EXCL dengan expected return harian 0,097%, risiko 1,530%, dan Sharpe Ratio 0,050, menunjukkan portofolio mampu mengurangi risiko sekaligus mendukung investasi berkelanjutan. Model peramalan berdasarkan hyperparameters terbak terbukti akurat dengan MAPE ≤10% pada seluruh saham, di mana BRIS memberikan imbal hasil tertinggi dengan risiko sangat kecil. Perbandingan kinerja portofolio menunjukkan bahwa model peramalan cukup baik dalam memproyeksikan expected return.
=================================================================================================================================
The capital market plays a crucial role in supporting economic growth, including in Indonesia. Public preference for sharia-compliant stocks has increased due to their relatively stable performance. One of the leading indices is the Jakarta Islamic Index (JII), which consists of 30 highly liquid sharia stocks. However, price fluctuations and investment risks remain inevitable, making diversification through optimal portfolio formation essential. This study applies Particle Swarm Optimization (PSO) as a metaheuristic approach to portfolio optimization by maximizing the Sharpe Ratio, combined with the Gated Recurrent Unit (GRU) model for forecasting. The results show that the optimal portfolio is formed through risk-free return and ESG Risk Rating screening, leaving only stocks with sustainable and promising prospects. The candidate stocks have distinct profiles. AKRA offers stable returns with low volatility. BRIS provides high returns with higher volatility. EXCL delivers moderate returns with low risk. PGAS performs fairly well with moderate risk. PGEO yields high returns but with greater risk. Their relatively low correlations indicate diversification opportunities.PSO optimization produces four weighted stocks in the optimal portfolio, namely PGAS, BRIS, PGEO, and EXCL, with an expected daily return of 0,097%, risk of 1,530%, and a Sharpe Ratio of 0,050. This highlights the portfolio’s ability to reduce risk while supporting sustainable investment. The GRU model shows strong accuracy with MAPE not exceeding 10%. BRIS records the highest return with minimal risk. Portfolio performance comparison confirms the forecasting model is valid in projecting expected returns.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Gated Recurrent Unit, Jakarta Islamic Index, Particle Swarm Optimization, Peramalan, Potofolio Optimal Forecasting, Gated Recurrent Unit, Jakarta Islamic Index, Optimal Portfolio, Particle Swarm Optimization
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > Q Science (General) > Q337.3 Swarm intelligence Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Amanda Shofiyah Salsabila
Date Deposited:	29 Jan 2026 05:14
Last Modified:	29 Jan 2026 05:14
URI:	http://repository.its.ac.id/id/eprint/130968

Actions (login required)

View Item