Optimasi Portofolio Markowitz dan Peramalan Harga Saham GRU-Sentimen pada Indeks IDX High Dividend 20

Hikmawati, Alifa (2026) Optimasi Portofolio Markowitz dan Peramalan Harga Saham GRU-Sentimen pada Indeks IDX High Dividend 20. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5002221165-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (16MB) | Request a copy

Abstract

Pergerakan harga saham yang tidak menentu memerlukan strategi investasi yang mampu mengelola risiko sekaligus menghasilkan prediksi yang akurat. Penelitian ini bertujuan membentuk portofolio optimal saham indeks IDX High Dividend 20 menggunakan model Markowitz serta menganalisis pengaruh integrasi sentimen berita keuangan terhadap kinerja peramalan harga saham menggunakan Gated Recurrent Unit (GRU). Sentimen berita diekstraksi menggunakan FinBERT dan digunakan sebagai fitur tambahan dalam proses peramalan. Evaluasi dilakukan menggunakan metrik Mean Absolute Percentage Error (MAPE) dan Root Mean Squared Error (RMSE) pada enam skema pengujian dengan variasi sequence length dan horizon peramalan. Hasil optimasi menghasilkan bobot investasi optimal sebesar 6,11% untuk ACES, 32,21% untuk BBRI, 23,63% untuk BMRI, dan 38,05% untuk BNGA. Hasil peramalan menunjukkan bahwa integrasi sentimen berita mampu meningkatkan kinerja model dibandingkan model tanpa sentimen. Skema terbaik diperoleh pada sequence length 60 hari dan horizon 7 hari dengan MAPE rata-rata sebesar 2,795% dan RMSE rata-rata sebesar Rp100,30. Evaluasi prospek portofolio menggunakan compound return menunjukkan bahwa hasil peramalan pada skema tersebut hanya berselisih 0,1681% dari compound return aktualnya. Hasil penelitian menunjukkan bahwa kombinasi optimasi portofolio Markowitz dan model GRU berbasis sentimen dapat digunakan untuk mendukung pengambilan keputusan investasi yang lebih efektif.
===================================================================================================================================
Unpredictable stock price movements require investment strategies that can manage risk while producing accurate predictions. This study aims to construct an optimal portfolio of IDX High Dividend 20 stocks using the Markowitz model and to analyze the effect of integrating financial news sentiment into a Gated Recurrent Unit (GRU) model for stock price forecasting. News sentiment was extracted using FinBERT and incorporated as an additional feature in the forecasting process. Model performance was evaluated using Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE) across six experimental schemes with different sequence lengths and forecasting horizons. The portfolio optimization results produced optimal investment weights of 6.11% for ACES, 32.21% for BBRI, 23.63% for BMRI, and 38.05% for BNGA. The forecasting results show that integrating news sentiment improves model performance compared to the baseline model without sentiment. The best performance was achieved with a sequence length of 60 days and a forecasting horizon of 7 days, resulting in an average MAPE of 2.795% and an average RMSE of Rp100.30. Portfolio prospect evaluation using compound return shows that the forecast from this scheme differs from the actual compound return by only 0.1681%. These findings indicate that the combination of Markowitz portfolio optimization and sentiment-enhanced GRU forecasting can support more effective investment decision-making.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Optimasi Portofolio, Markowitz, GRU, FinBERT, Analisis Sentimen, Peramalan Harga Saham, Portfolio Optimization, Markowitz, GRU, FinBERT, Sentiment Analysis, Stock Price Forecasting.
Subjects:	Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Alifa Hikmawati
Date Deposited:	29 Jul 2026 01:16
Last Modified:	29 Jul 2026 01:16
URI:	http://repository.its.ac.id/id/eprint/139112

Actions (login required)

View Item