Maurita, Eka (2026) ADRO Stock Price Prediction Using Hybrid GRU - XGBoost. Undergraduate thesis, UPN Veteran Jawa Timur.
![[img]](https://repository.upnjatim.ac.id/style/images/fileicons/text.png) |
Text (Cover)
22081010119.-Cover.pdf Download (1MB) |
|
Text (Bab1)
22081010119.-bab1.pdf Download (284kB) |
|
Text (Bab2)
22081010119.-bab2.pdf Restricted to Repository staff only until 17 July 2028. Download (591kB) |
|
Text (Bab3)
22081010119.-bab3.pdf Restricted to Repository staff only until 17 July 2028. Download (722kB) |
|
Text (Bab4)
22081010119.bab4.pdf Restricted to Repository staff only until 17 July 2028. Download (1MB) |
|
Text (Bab5)
220810101119.-bab5.pdf Download (142kB) |
|
Text (Daftar Pustaka)
22081010119.-daftarpustaka.pdf Download (182kB) |
|
Text (Lampiran)
22081010119.-lampiran.pdf Restricted to Repository staff only Download (165kB) |
Abstract
PT AlamTri Resources Indonesia Tbk. (ADRO) shares are energy sector stocks that have high volatility and are influenced by external factors such as world coal prices and the USD/IDR exchange rate, making it interesting to study using an artificial intelligence-based prediction approach. This study aims to predict the stock price of PT AlamTri Resources Indonesia Tbk. (ADRO) using the Hybrid Gated Recurrent Unit (GRU) and Extreme Gradient Boosting (XGBoost) methods by utilizing historical data of ADRO stock prices, world coal prices, and the USD/IDR exchange rate for the 2018–2025 period. The research stages include data collection, preprocessing, correlation analysis, normalization, sequence data formation, GRU, XGBoost, and Hybrid GRU-XGBoost modeling, and evaluation using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE). The results showed that the Hybrid GRU-XGBoost model provided the best performance compared to the GRU and XGBoost models individually with an MAE of 36.24, RMSE of 55.03, and MAPE of 1.92%, while the GRU model produced an MAE of 38.56, RMSE of 57.55, and MAPE of 2.05%, and the XGBoost model produced an MAE of 51.87, RMSE of 67.19, and MAPE of 2.81%. These results indicate that the combination of GRU which is able to capture temporal patterns and XGBoost which is able to model non-linear relationships can improve the accuracy of ADRO stock price predictions. The resulting model was then implemented into a Streamlit-based application to support interactive visualization and analysis of prediction results. Keywords: ADRO, GRU, XGBoost, Hybrid GRU-XGBoost Stock Prices, Time Series Prediction.
| Item Type: | Thesis (Undergraduate) | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
||||||||||||
| Subjects: | T Technology > T Technology (General) > T58.6-58.62 Management Information Systems | ||||||||||||
| Divisions: | Faculty of Computer Science > Departemen of Informatics | ||||||||||||
| Depositing User: | Eka eka maurita | ||||||||||||
| Date Deposited: | 17 Jul 2026 06:56 | ||||||||||||
| Last Modified: | 17 Jul 2026 06:59 | ||||||||||||
| URI: | https://repository.upnjatim.ac.id/id/eprint/55768 |
Actions (login required)
 |
View Item |