Peramalan Arah Pergerakan Harga Saham Berdasarkan Indikator Teknikal dengan Pendekatan Random Forest dan Quantile Regression Forest

Soleha, Harista Almiatus (2025) Peramalan Arah Pergerakan Harga Saham Berdasarkan Indikator Teknikal dengan Pendekatan Random Forest dan Quantile Regression Forest. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6003231008-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2027. Download (5MB) | Request a copy

Abstract

Investasi saham menjadi pilihan populer di kalangan investor karena menawarkan potensi keuntungan yang menarik. Salah satu sektor saham yang menarik adalah sektor perbankan, karena memiliki kapitalisasi pasar tertinggi. Penelitian ini menggunakan tiga saham yaitu BBCA, BBTN, dan ARTO. Pergerakan harga saham sering kali memiliki karakteristik perubahan yang cepat. Oleh karena itu, kemampuan untuk meramalkan pergerakan harga saham menjadi penting bagi investor dan perusahaan. Untuk meramalkan harga saham, beberapa model yang didasarkan pembelajaran mesin telah dikembangkan salah satunya yaitu Random Forest (RF). RF hanya memberikan prediksi rata-rata tanpa menunjukkan distribusi lengkap hasil prediksi. Untuk mengatasi hal tersebut, digunakan Quantile Regression Forest (QRF) yang tidak hanya menghasilkan nilai rata-rata, tetapi juga memberikan distribusi bersyarat dari setiap prediksi, memungkinkan analisis yang lebih detail. Penelitian ini bertujuan untuk melakukan kajian algoritma QRF pada data time seires dan mengetahui bagaimana peramalan arah pergerakan harga saham berdasarkan indikator teknikal menggunakan pendekatan RF dan QRF. Pada analisis RF harga saham diklasifikasikan menjadi dua kelas yaitu naik (1) dan turun (-1), sedangkan QRF akan membentuk interval, dan akan menggunakan horizon 5, 10, dan 22 hari perdagangan. Hasil dari analisis diperoleh bahwa QRF dapat diterapkan pada data time series dengan menggunakan block bootstrap, analisis RF diperoleh bahwa peramalan saham BBCA paling baik diramalkan pada horizon 10 hari, sedangkan BBTN dan ARTO pada horizon 22 hari. QRF saham BBCA menghasilkan nilai kesalahan terkecil dibandingkan dengan BBTN dan ARTO. Selain itu pada interval 90% saham BBCA memiliki interval peramalan paling sempit, diikuti oleh BBTN dan ARTO yang lebih lebar.
===================================================================================================================================
Stock investment is a popular choice among investors as it offers attractive profit potential. One interesting stock sector is the banking sector, because it has the highest market capitalization. This study uses three stocks, namely BBCA, BBTN, and ARTO. Stock price movements are often characterized by rapid changes. Therefore, the ability to forecast stock price movements is important for investors and companies. To forecast stock prices, several models based on machine learning have been developed, one of which is Random Forest (RF). RF only provides an average prediction without showing the complete distribution of the prediction results. To overcome this, Quantile Regression Forest (QRF) is used which not only produces an average value, but also provides the conditional distribution of each prediction, allowing for more detailed analysis. This research aims to conduct a study of the QRF algorithm on time series data and find out how to forecast the direction of stock price movements based on technical indicators using RF and QRF approaches. In RF analysis, stock prices are classified into two classes, namely up (1) and down (-1), while QRF will form an interval, and will use a horizon of 5, 10, and 22 trading days. The results of the analysis obtained that QRF can be applied to time series data using block bootstrap, RF analysis obtained that BBCA stock forecasting is best predicted at a 10-day horizon, while BBTN and ARTO at a 22-day horizon. QRF of BBCA shares produces the smallest error value compared to BBTN and ARTO. In addition, at 90% interval, BBCA shares have the narrowest forecasting interval, followed by BBTN and ARTO which are wider.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Saham, Random Forest, Quantile Regression Forest, Time Series Stock, Random Forest, Quantile Regression Forest, Time Series
Subjects:	H Social Sciences > HA Statistics > HA30.3 Time-series analysis H Social Sciences > HG Finance H Social Sciences > HG Finance > HG4915 Stocks--Prices Q Science Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49101-(S2) Master Thesis
Depositing User:	Harista Almiatus Soleha
Date Deposited:	06 Feb 2025 03:55
Last Modified:	06 Feb 2025 03:55
URI:	http://repository.its.ac.id/id/eprint/118341

Actions (login required)

View Item