PREDIKSI INDEKS HARGA SAHAM GABUNGAN (IHSG) MENGGUNAKAN MODEL HYBRID EXPONENTIAL GENERALIZED AUTOREGRESSIVE CONDITIONAL HETEROSCEDASTICITY-LONG SHORT TERM MEMORY (EGARCH-LSTM)

    Gabriella Martha Wolff, - and Dewi Rachmatin, - and Fitriani Agustina, - (2025) PREDIKSI INDEKS HARGA SAHAM GABUNGAN (IHSG) MENGGUNAKAN MODEL HYBRID EXPONENTIAL GENERALIZED AUTOREGRESSIVE CONDITIONAL HETEROSCEDASTICITY-LONG SHORT TERM MEMORY (EGARCH-LSTM). S1 thesis, Universitas Pendidikan Indonesia.

    Abstract

    Salah satu jenis investasi yang paling populer saat ini adalah investasi saham. Indeks Harga Saham Gabungan (IHSG) dapat menjadi salah satu indikator penting yang perlu diperhatikan bagi calon investor saham dalam mengambil keputusan di pasar modal Indonesia. IHSG menggambarkan pergerakan fluktuatif harga saham secara umum di Indonesia. Dalam penelitian ini akan memprediksi nilai IHSG dengan memodelkan data historis hari IHSG periode Januari 2022-April 2025 dengan model hybrid EGARCH-LSTM. Model EGARCH secara khusus akan menangani efek heteroskedastisitas dan asimetris pada data keuangan, serta pendekatan neural network dengan model LSTM akan dipergunakan untuk menangkap pola nonlinear dari residu model EGARCH. Pada penelitian ini, data akan dibagi menjadi 85% untuk data latih dan 15% untuk data uji. Berdasarkan hasil pengolahan data menggunakan software RStudio dan Python, diperoleh model terbaik, yaitu EGARCH(1,1) dan LSTM dengan kombinasi parameter window size 30, time step 30, learning rate 0,001, batch size 16, neurons 128, dan epochs 100, serta menggunakan optimizer Adam. Model hybrid EGARCH-LSTM mampu memberikan hasil prediksi IHSG untuk lima periode ke depan hasil yang lebih baik dibanding dengan model EGARCH(1,1) dengan akurasi yang sangat baik, ditunjukkan oleh nilai RMSE sebesar 66,837 dan MAPE sebesar 0,713%, yang berada jauh di bawah ambang batas 10% untuk kategori akurasi prediksi sangat baik. Kata Kunci: Saham, Prediksi, IHSG, EGARCH, LSTM One of the most popular types of investment today is stock investment. An important indicator that prospective investors should consider when making decisions in the Indonesian capital market is the Indeks Harga Saham Gabungan (IHSG), which reflects the overall fluctuation of stock prices in Indonesia. Therefore, this study aims to predict the value of the IHSG by modeling daily historical IHSG data from January 2022 to April 2025 using a hybrid EGARCH-LSTM model. The EGARCH model is specifically employed to address heteroskedasticity and asymmetry in financial data, while the neural network approach using the LSTM model is used to capture nonlinear patterns from the residuals of the EGARCH model. In this study, the data is divided into 85% training data and 15% testing data. Based on data processing using RStudio and Python, the best-performing model was obtained: EGARCH(1,1) and LSTM with a parameter combination of window size 30, time step 30, learning rate 0,001, batch size 16, neurons 128, and epochs 100, using the Adam optimizer. The hybrid EGARCH-LSTM model successfully provided better IHSG predictions for the next five periods compared to the EGARCH(1,1) model alone, with excellent accuracy, as indicated by an RMSE value of 66,837 and a MAPE of 0,713%, which is well below the 10% threshold for excellent prediction accuracy. Keywords: Stocks, Prediction, IHSG, EGARCH, LSTM

    Baca Full Text klik disini
    [thumbnail of S_MAT_2100159_Title.pdf] Text
    S_MAT_2100159_Title.pdf - Cover Image
    Download (467kB)
    [thumbnail of S_MAT_2100159_Chapter1.pdf] Text
    S_MAT_2100159_Chapter1.pdf
    Download (61kB)
    [thumbnail of S_MAT_2100159_Chapter2.pdf] Text
    S_MAT_2100159_Chapter2.pdf
    Restricted to Staf Perpustakaan
    Download (527kB)
    [thumbnail of S_MAT_2100159_Chapter3.pdf] Text
    S_MAT_2100159_Chapter3.pdf
    Download (328kB)
    [thumbnail of S_MAT_2100159_Chapter4.pdf] Text
    S_MAT_2100159_Chapter4.pdf
    Restricted to Staf Perpustakaan
    Download (753kB)
    [thumbnail of S_MAT_2100159_Chapter5.pdf] Text
    S_MAT_2100159_Chapter5.pdf
    Download (85kB)
    [thumbnail of S_MAT_2100159_Appendix.pdf] Text
    S_MAT_2100159_Appendix.pdf
    Restricted to Staf Perpustakaan
    Download (1MB)
    Official URL: https://repository.upi.edu/
    Item Type: Thesis (S1)
    Additional Information: ID SINTA Dosen Pembimbing : Dewi Rachmatin: 5975775 Fitriani Agustina: 5981275
    Uncontrolled Keywords: Saham, Prediksi, IHSG, EGARCH, LSTM
    Subjects: Q Science > QA Mathematics
    Divisions: Fakultas Pendidikan Matematika dan Ilmu Pengetahuan Alam > Program Studi Matematika - S1 > Program Studi Matematika (non kependidikan)
    Depositing User: Gabriella Martha Wolff
    Date Deposited: 11 Sep 2025 04:16
    Last Modified: 11 Sep 2025 04:16
    URI: http://repository.upi.edu/id/eprint/138277

    Actions (login required)

    View Item View Item