cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Andhika Rafi Hananto
Contact Email
andhikarh90@gmail.com
Phone
+6282314736799
Journal Mail Official
support@jdmdc.com
Editorial Address
Graha Permata Estate, Jl. HM Bahrun Blok H9, Sokayasa, Berkoh, Kec. Purwokerto Tim., Kabupaten Banyumas, Jawa Tengah 53146
Location
Kab. banyumas,
Jawa tengah
INDONESIA
Journal of Digital Market and Digital Currency
Published by Meta Bright Indonesia
ISSN : -     EISSN : 30480981     DOI : https://doi.org/10.47738/jdmdc
Core Subject : Economy, Science,
Computer Science & IT Decision Sciences, Operations Research & Management Economics, Econometrics & Finance
Journal of Digital Market and Digital Currency publishes high-quality research on: Digital Marketing Digital Currencies Cryptocurrency Trends Blockchain Applications Fintech Innovations Our goal is to provide a platform for researchers, practitioners, and policymakers to share innovative findings, discuss emerging trends, and address the challenges and opportunities presented by the Journal of Digital Market and Digital Currency.
Arjuna Subject : Bisnis, Menejemen, dan Akutansi (semua kategori) - Bisnis, Menejemen, dan Akutansi (Lain-Lain)
Articles 50 Documents
 1   2   3   4   5 
Machine Learning-Based Forecasting of AAVE Cryptocurrency: A Comparative Study of Regression, Ensemble, and Deep Learning Models Monika Łobaziewicz
Journal of Digital Market and Digital Currency Vol. 3 No. 2 (2026): Regular Issue June 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jdmdc.v3i2.62

Abstract

The volatility of cryptocurrency markets has increased the demand for accurate forecasting models that can help investors and analysts anticipate price movements. This study evaluates the predictive performance of four machine learning algorithms, namely Linear Regression, Random Forest, XGBoost, and Long Short-Term Memory (LSTM), in forecasting the closing price of the AAVE cryptocurrency. The models were trained using historical market data consisting of key indicators such as Open, High, Low, Volume, and Marketcap. Their performance was assessed using Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and the Coefficient of Determination (R²). The results show that Linear Regression produced the most accurate predictions with the lowest MAE (8.13), RMSE (8.76), and the highest R² (0.9924). Random Forest and XGBoost also achieved good results with R² values of 0.9337 and 0.9484, respectively, while the LSTM model performed poorly with an R² of 0.4328. The study concludes that simpler models can outperform more complex algorithms when the dataset is limited and exhibits linear behavior. The findings emphasize that model selection in cryptocurrency forecasting should consider data structure and quantity. Future work should involve larger datasets, higher-frequency data, and hybrid models that integrate ensemble learning and deep learning for improved predictive accuracy.
A Hybrid SARIMAX–LSTM Framework for Predicting Price Volatility in High-Tech Digital Markets: Evidence from NVIDIA Flurey Martin
Journal of Digital Market and Digital Currency Vol. 3 No. 2 (2026): Regular Issue June 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jdmdc.v3i2.63

Abstract

This study develops a Hybrid SARIMAX–LSTM model to improve the accuracy and robustness of stock price forecasting in digital and volatile financial markets. The model combines the linear and seasonal forecasting strengths of the Seasonal Autoregressive Integrated Moving Average with Exogenous Variables (SARIMAX) with the nonlinear learning capability of the Long Short-Term Memory (LSTM) network. Using historical data from NVDIA Corporation, the hybrid framework was optimized through smart weighting to balance the contribution of both components. The results show that the model achieved a Root Mean Square Error (RMSE) of 8.59 and a coefficient of determination (R²) of 0.9166, indicating that over 91 percent of price variance was accurately explained. Residual analysis confirmed unbiased predictions with normally distributed errors, demonstrating high stability and adaptability under volatile market conditions. Compared with individual models, the hybrid approach produced smoother and more consistent forecasts. Overall, the Hybrid SARIMAX–LSTM framework offers an interpretable and reliable tool for digital market forecasting and AI-based financial decision-making.
A Reinforcement Learning Approach to Bitcoin Trading: Proximal Policy Optimization with Trend-Following and Risk-Aware Reward Design Victor Vladareanu
Journal of Digital Market and Digital Currency Vol. 3 No. 2 (2026): Regular Issue June 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jdmdc.v3i2.64

Abstract

This study proposes a reinforcement learning based trading strategy for Bitcoin using Proximal Policy Optimization with a trend following and risk aware reward design. The model is developed within a custom trading environment that incorporates multiple technical indicators, including trend, momentum, and volatility features, to capture market dynamics. A continuous action space is employed to enable flexible portfolio allocation between cash and Bitcoin, allowing the agent to learn dynamic position sizing rather than discrete buy or sell decisions. The reward function is designed to encourage profit generation while penalizing excessive risk, trading activity, and drawdowns. The proposed model is evaluated on historical Bitcoin data and compared with a Buy and Hold baseline using metrics such as total return, Sharpe ratio, maximum drawdown, trading frequency, and transaction costs. The results show that while the PPO strategy does not outperform Buy and Hold in terms of total return, it achieves superior risk adjusted performance with a higher Sharpe ratio and more stable portfolio growth. However, the model exhibits high trading frequency, leading to increased transaction costs that reduce overall profitability. These findings demonstrate that reinforcement learning offers a promising approach for developing adaptive and risk sensitive trading strategies, although further improvements are required to enhance trading efficiency and cost management.
Enhancing Loan Approval Prediction Using Ensemble Machine Learning Techniques Through Comprehensive Model Comparison and Performance Evaluation Analysis Qing Tan
Journal of Digital Market and Digital Currency Vol. 3 No. 2 (2026): Regular Issue June 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jdmdc.v3i2.65

Abstract

Loan approval prediction is a crucial task in the financial sector, as it directly impacts risk management and decision-making processes. This study aims to enhance the accuracy of loan approval prediction by applying ensemble machine learning techniques and comparing their performance with a baseline model. The dataset used in this study contains borrower demographic, financial, and employment-related attributes, and missing values were handled using a deletion method to ensure data consistency. Several models were implemented, including Logistic Regression as the baseline model, as well as ensemble methods such as Random Forest, Gradient Boosting, and Voting Classifier. The models were evaluated using multiple performance metrics, including Accuracy, Precision, Recall, F1-Score, and ROC-AUC. The experimental results show that ensemble models consistently outperform the baseline model across all evaluation metrics. Random Forest achieved the highest ROC-AUC, indicating superior discriminative capability, while the Voting Classifier provided the best balance between precision and recall, resulting in the highest F1-Score. In addition, feature importance analysis revealed that CreditScore, Income, and Employment Type are the most influential factors in loan approval decisions. These findings demonstrate that ensemble learning methods are effective in improving predictive performance and can provide reliable support for loan approval decision-making systems.
Machine Learning-Based Fraud Detection in Banking Transactions Using Integrated Behavioral and Transactional Feature Engineering with Weak Labeling Approach Hongbo Wang
Journal of Digital Market and Digital Currency Vol. 3 No. 2 (2026): Regular Issue June 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jdmdc.v3i2.66

Abstract

Fraud detection in banking transactions is a critical challenge due to the imbalanced nature of data and the lack of labeled fraud instances. This study proposes a machine learning approach for detecting fraudulent transactions by integrating behavioral and transactional features, combined with a rule-based weak labeling strategy to generate fraud labels. The dataset consists of 2,512 banking transactions, with 14.29% labeled as fraud. Three models were evaluated, including Logistic Regression, Random Forest, and XGBoost, using stratified cross-validation and multiple evaluation metrics such as accuracy, precision, recall, F1-score, and ROC-AUC. The results show that ensemble-based models outperform Logistic Regression, with Random Forest achieving the best balance between precision and recall, and XGBoost obtaining perfect recall and the highest ROC-AUC, indicating its strong ability to detect fraudulent transactions. Feature importance analysis reveals that transaction amount and deviation from typical user behavior are key indicators of fraud. Despite these promising results, the study is limited by the use of rule-based labeling and a relatively small dataset. Future work should focus on validating the proposed approach using real-world labeled data and improving model robustness for practical deployment.
Deep Learning-Based Sentiment Analysis for Digital Market Intelligence through Comparative Evaluation of BiLSTM and CNN Models Hendro Budiyanto; Muhammad Shahid Khan; Winda Widyanty
Journal of Digital Market and Digital Currency Vol. 3 No. 3 (2026): Regular Issue September 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jdmdc.v3i3.72

Abstract

The rapid growth of digital communication platforms has generated vast amounts of textual data containing valuable insights into public sentiment toward products, services, and digital assets. Accurately analyzing this data is crucial for understanding consumer behavior and market trends in digital marketing and cryptocurrency ecosystems. This study presents a comparative analysis of two deep learning architectures, Bidirectional Long Short-Term Memory (BiLSTM) and Convolutional Neural Network (CNN), for sentiment classification using textual data. Both models were trained for twenty epochs without early stopping to evaluate their full learning potential. The experimental results revealed that the BiLSTM model achieved superior performance with an accuracy of 73.82% and an F1-score of 72.40%, while the CNN model obtained 65.10% accuracy and an F1-score of 64.28%. The BiLSTM demonstrated a higher capability to capture sequential dependencies and contextual semantics through its bidirectional processing, whereas the CNN primarily relied on local feature extraction, limiting its contextual understanding. Class-wise analysis showed that BiLSTM performed strongly in identifying neutral and positive sentiments but struggled with negative sentiment detection due to data imbalance and linguistic ambiguity. These findings highlight BiLSTM’s robustness and suitability for sentiment analysis in digital market applications. The study emphasizes the importance of context-aware deep learning models in supporting data-driven marketing strategies, customer sentiment monitoring, and digital asset analysis. Future research is recommended to integrate transformer-based architectures such as BERT or RoBERTa to further enhance contextual comprehension and improve overall classification performance.
Explainable Ensemble Learning for Loan Approval Prediction Using XGBoost, LightGBM, and Random Forest with SHAP Analysis Mikaria Gultom; Chininta Rizka Angelia; Vijayaletchumy Krishnan
Journal of Digital Market and Digital Currency Vol. 3 No. 3 (2026): Regular Issue September 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jdmdc.v3i3.74

Abstract

Loan approval prediction is a critical task in financial institutions, as it directly impacts risk management and decision-making processes. However, challenges such as class imbalance and lack of model interpretability often limit the effectiveness and reliability of machine learning approaches. This study proposes an explainable ensemble learning framework for predicting loan approval using Random Forest, XGBoost, and LightGBM models. To address the issue of imbalanced data, the Synthetic Minority Over-sampling Technique (SMOTE) was applied to the training dataset, ensuring a balanced class distribution and improving model performance. The models were evaluated using multiple metrics, including Accuracy, Precision, Recall, F1-score, and ROC-AUC. The results show that all models achieved strong performance, with Random Forest outperforming the others by achieving the highest accuracy of 0.959 and ROC-AUC of 0.9386. Furthermore, SHapley Additive exPlanations (SHAP) were employed to enhance model interpretability by providing both global and local explanations of feature contributions. The analysis revealed that CreditScore, EmploymentType, and Income are the most influential factors in determining loan approval decisions. In addition, SHAP dependence and interaction analysis demonstrated the presence of non-linear relationships and feature interactions, offering deeper insight into the model’s decision-making process.
Prediction of Tesla Stock Prices Using Recurrent Neural Networks (RNN) within Financial Technology (FinTech) Framework Andy Firmansyah; Mochammad Fahlevi; Daniel Wisuda Purba
Journal of Digital Market and Digital Currency Vol. 3 No. 3 (2026): Regular Issue September 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jdmdc.v3i3.88

Abstract

The integration of artificial intelligence (AI) into Financial Technology (FinTech) has transformed the landscape of financial data analysis and prediction. This research presents a predictive model for Tesla stock prices using a Recurrent Neural Network (RNN) as part of an AI-based FinTech framework. Historical Tesla stock data obtained from the Nasdaq market were normalized and structured using a 60-day time window to train the RNN model. The results show that the proposed model achieved a prediction accuracy of 93.45 percent, with a Mean Absolute Error (MAE) of 43.14, Mean Squared Error (MSE) of 3148.46, and Root Mean Squared Error (RMSE) of 56.11. The convergence of training and validation losses confirmed the model’s stability, while residual analysis indicated unbiased predictions centered around zero. These findings demonstrate that the RNN effectively captures temporal dependencies in stock price time-series data. Although the Long Short-Term Memory (LSTM) model achieved a slightly higher accuracy of 94.82 percent, the RNN remains advantageous due to its lower computational cost and faster training process. Overall, the RNN model offers a reliable and efficient approach for real-time FinTech applications such as automated trading systems, portfolio optimization, and investment risk analysis.
Hybrid CNN-LSTM Deep Learning Framework for Predicting Stock Market Movements Using Technical Indicators and Historical Price Data Jihad Abou Sondos
Journal of Digital Market and Digital Currency Vol. 3 No. 3 (2026): Regular Issue September 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jdmdc.v3i3.92

Abstract

This study investigates the application of deep learning models for stock market forecasting, focusing on the comparative performance of a baseline Long Short-Term Memory (LSTM) model and a Hybrid Convolutional Neural Network–Long Short-Term Memory (CNN-LSTM) architecture. The research aims to evaluate whether the integration of convolutional and recurrent layers can enhance predictive accuracy, stability, and generalization in financial time-series modeling. Historical stock data from Apple Inc. were used, with 80 percent allocated for training and 20 percent for validation. The models were trained using technical indicators and closing price data, and their performance was evaluated using Mean Squared Error (MSE), Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE). The results show that both models achieved high forecasting accuracy, with RMSE values below 40, demonstrating strong capability in capturing nonlinear market dynamics. Although the LSTM-only model achieved slightly lower numerical error values, the Hybrid CNN-LSTM produced smoother and more stable predictions with better resistance to market noise and short-term volatility. The residual distribution and convergence patterns confirmed that the hybrid model achieved balanced performance and effective generalization. These findings indicate that the Hybrid CNN-LSTM framework provides a reliable approach for stock market forecasting, offering both precision and stability. The study concludes that this hybrid deep learning approach holds significant potential for supporting digital financial analytics and long-term investment decision-making.
Comparative Study of Deep Learning Architectures for Financial Sentiment Analysis in Digital Market Environments Lakshmi Dandapani; Manimegalai Subramanian
Journal of Digital Market and Digital Currency Vol. 3 No. 3 (2026): Regular Issue September 2026
Publisher : Bright Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47738/jdmdc.v3i3.83

Abstract

This study investigates the effectiveness of deep learning architectures for financial sentiment classification using textual data from financial reports and market statements. Four models, namely Long Short-Term Memory (LSTM), Bidirectional LSTM (BiLSTM), Gated Recurrent Unit (GRU), and Convolutional Neural Network (CNN), were trained and evaluated on a dataset of 5,842 labeled financial sentences categorized as positive, negative, or neutral. The evaluation employed accuracy and weighted F1-score metrics to assess predictive performance and class balance. The results revealed that the CNN model achieved the highest performance with an accuracy of 70.32 percent and an F1-score of 70.22 percent, followed by the BiLSTM model with 67.49 percent accuracy and 66.86 percent F1-score. The CNN’s superior performance demonstrates its ability to capture local sentiment-bearing features efficiently, while BiLSTM excels in understanding bidirectional contextual dependencies. In contrast, the LSTM and GRU models performed less effectively due to their reliance on sequential memory mechanisms, which are less applicable to short and factual financial sentences. These findings highlight that CNN provides the best balance between accuracy, efficiency, and adaptability for financial sentiment classification and can serve as a foundation for intelligent digital market analysis systems and AI-driven financial decision support tools.

Page 5 of 5 | Total Record : 50

 1   2   3   4   5 

Filter by Year

2024 2026


Reset
Filter By Issues
All Issue Vol. 3 No. 3 (2026): Regular Issue September 2026 Vol. 3 No. 2 (2026): Regular Issue June 2026 Vol. 3 No. 1 (2026): Regular Issue March 2026 Vol. 2 No. 4 (2025): Regular Issue December 2025 Vol. 2 No. 3 (2025): Regular Issue September 2025 Vol. 2 No. 2 (2025): Regular Issue June 2025 Vol. 2 No. 1 (2025): Regular Issue March 2025 Vol. 1 No. 3 (2024): Regular Issue December 2024 Vol. 1 No. 2 (2024): Regular Issue September 2024 Vol. 1 No. 1 (2024): Regular Issue June 2024