Price Prediction and Stock Buying and Selling Rates Using Deep Learning and Deep Q Network

Luna, Nabila Anggita (2026) Price Prediction and Stock Buying and Selling Rates Using Deep Learning and Deep Q Network. Undergraduate thesis, UPN Veteran Jawa Timur.

[img] Text
cover.pdf

Download (3MB)
[img] Text
BAB 1.pdf

Download (840kB)
[img] Text
BAB 2.pdf
Restricted to Repository staff only until 12 November 2028.

Download (1MB)
[img] Text
bab 3.pdf
Restricted to Repository staff only until 12 November 2028.

Download (1MB)
[img] Text
BAB 4.pdf
Restricted to Repository staff only until 12 November 2028.

Download (3MB)
[img] Text
bab 5.pdf

Download (626kB)
[img] Text
Daftar pustaka.pdf

Download (365kB)

Abstract

This study aims to develop a hybrid BiGRU-LSTM model for stock price forecasting and integrate forecasting results into the Deep Q-Network (DQN) algorithm as an investment decision-making system. The research object consists of three energy sector issuers, namely ADRO, BUMI, and PTRO. The BiGRU LSTM model is used to study the temporal patterns of stock prices, while DQN is used to generate Buy, Hold, and Sell decisions. The test results showed that the BiGRU-LSTM model was able to produce good forecasting performance in all three issuers with an R² value above 0.92. ADRO obtained an RMSE value of 71.89, MAE 53.37, R² 0.9602, and MAPE of 2.69%; BUMI obtained RMSE 17.81, MAE 10.59, R² 0.9607, and MAPE 4.44%; while PTRO obtained RMSE 716.00, MAE 531.00, R² 0.92, and MAPE 8.02%. Based on MAPE as the main metric, ADRO has the best forecasting performance. Furthermore, the integration of prediction results with DQN resulted in positive returns in all three issuers, namely 6.13% in BUMI, 16.89% in PTRO, and 6.88% in ADRO, and resulted in a lower risk of portfolio decline compared to the Buy and Hold strategy. The results of the study show that the combination of BiGRU-LSTM and DQN can be used as an approach for price forecasting while supporting stock investment decision-making. Keywords: Deep Learning, BiGRU-LSTM, Deep Q-Network, Stock Price Prediction, Energy Sector, Reinforcement Learning.

Item Type: Thesis (Undergraduate)
Contributors:
ContributionContributorsNIDN/NIDKEmail
Thesis advisorMade Hanindia Prami, Swari0805028901madehanindia.fik@upnjatim.ac.id
Thesis advisorRizky, Parlika0718058401rizkyparlika.if@upnjatim.ac.id
Subjects: Q Science > QA Mathematics > QA76 Computer software
Q Science > QA Mathematics > QA76.6 Computer Programming
T Technology > TN Mining engineering. Metallurgy
Divisions: Faculty of Computer Science > Departemen of Informatics
Depositing User: Nabila anggita luna
Date Deposited: 13 Nov 2026 03:20
Last Modified: 13 Nov 2026 03:20
URI: https://repository.upnjatim.ac.id/id/eprint/61329

Actions (login required)

View Item View Item