Riskullah, Faiz (2023) Analisa Pengaruh Faktor Kurs Dollar, Komoditas, Dan Bursa Asing Terhadap IHSG Dengan Metode Supervised Machine Learning. Masters thesis, Institut Teknologi Sepuluh Nopember.
Text
6032202105-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2025. Download (3MB) | Request a copy |
Abstract
Indonesia merupakan negara berkembang yang memiliki pertumbuhan ekonomi pesat sehingga memberikan citra positif bagi para investor untuk berinvestasi. Salah satu instrumen keuangan yang diminati saat ini yaitu saham. Hal ini dibuktikan dengan kinerja IHSG selama 20 tahun terakhir mengalami peningkatan sebesar 15,15%/tahun. Kondisi tersebut tidak hanya menarik investor asing tetapi investor dalam negeri. Beberapa faktor yang mempengaruhi pergerakan IHSG diantaranya komoditas, kondisi ekonomi dalam negeri, pergerakan indeks bursa asing, dan lain-lain. Pada penelitian ini dilakukan analisa korelasi untuk memberikan gambaran terhadap faktor yang mempengaruhi pergerakan IHSG, dilanjutkan dengan pengolahan data faktor-faktor yang mempengaruhi IHSG untuk memprediksi pergerakan IHSG pada hari berikutnya dengan menggunakan metode supervised machine learning diantaranya Linear Regression, Gradient Boosted, Random Forest, dan Tree Ensemble. Hasil analisa korelasi pada penelitian ini menunjukan bahwa faktor bursa asing yang terdiri dari indeks Dow jones, indeks Hangseng, indeks Kospi, indeks Nikkei, indeks Nasdaq serta nilai tukar dollar berpengaruh positif terhadap IHSG, kemudian faktor komoditas yang terdiri dari crude oil, coal, dan gold juga mememiliki pengaruh positif terhadap IHSG. Sedangkan, sisanya yaitu crude palm oil berpengaruh negatif dengan nilai yang kecil. Hasil evaluasi dari 4 metode yang digunakan untuk prediksi nilai aktual IHSG dengan menggunakan parameter MAE, RMSE, dan MAPE menunjukkan bahwa metode tree ensemble merupakan metode yang tepat pada penelitian ini. Tekahir, uji deployment dilakukan dengan model yang telah didapatkan dan menghasilkan nilai MAE, RSME, dan MAPE sebesar 258,8780; 279,1176; dan 0,0364. Nilai tersebut lebih besar dibandingkan saat pemilihan metode yaitu sebesar 45,5301; 63,2720; dan 0,0082
=====================================================================================================================================
Indonesia is a developing country that has rapid economic growth which gives a positive image for investors to invest. One of the most popular financial instruments today is stocks. This is evidenced by the performance of the JCI over the last 20 years, which has increased by 15.15%/year. These conditions not only attract foreign investors but domestic investors. Some of the factors that influence the movement of the JCI include commodities, domestic economic conditions, movements in foreign stock indices, and others. In this study, correlation analysis was carried out to provide an overview of the factors that influence the movement of the JCI, followed by data processing of the factors that influence the JCI to predict the movement of the JCI the following day using supervised machine learning methods including Linear Regression, Gradient Boosted, Random Forest, and TreeEnsemble. The results of the correlation analysis in this study show that foreign exchange factors consisting of the Dow Jones index, Hangseng index, Kospi index, Nikkei index, Nasdaq index and the dollar exchange rate have a positive effect on the JCI, then commodity factors consisting of crude oil, coal, and gold also has a positive influence on the JCI. Meanwhile, the rest, namely crude palm oil, has a negative effect with a small value. The evaluation results of the 4 methods used to predict the actual value of the JCI using the MAE, RMSE, and MAPE parameters show that the tree ensemble method is the right method in this study. Finally, the deployment test was carried out with the model that was obtained and resulted in MAE, RSME, and MAPE values of 258.8780; 279.1176; and 0.0364. This value is greater than when selecting the method, which is equal to 45.5301; 63.2720; and 0.0082
Item Type: | Thesis (Masters) |
---|---|
Uncontrolled Keywords: | IHSG, Korelasi, Saham, Tree Ensemble, Correlation, JCI, Stocks |
Subjects: | H Social Sciences > HA Statistics > HA31.3 Regression. Correlation |
Divisions: | Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT) |
Depositing User: | Faiz Riskullah |
Date Deposited: | 12 Feb 2023 15:52 |
Last Modified: | 13 Feb 2023 01:21 |
URI: | http://repository.its.ac.id/id/eprint/97114 |
Actions (login required)
View Item |