Article Per Year (5 Year)
p-Index From 2021 - 2026
0.778
P-Index
This Author published in this journals
All Journal IAES International Journal of Artificial Intelligence (IJ-AI) JUITA : Jurnal Informatika Jurnal ULTIMATICS Journal of Information Systems and Informatics
Aditiya Hermawan
Buddhi Dharma University
Computer Science & IT Control & Systems Engineering Electrical & Electronics Engineering Engineering

Published : 4 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 4 Documents
Search

 1 
Optimizing Sentiment Classification of E-Commerce Product Reviews: A Comparative Study of Naïve Bayes and SVM with SMO Riki; Sonya Eliesse Dameria; Aditiya Hermawan; Junaedi; Yusuf Kurnia
JUITA: Jurnal Informatika JUITA Vol. 13 Issue 3, November 2025
Publisher : Department of Informatics Engineering, Universitas Muhammadiyah Purwokerto

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.30595/juita.v13i3.26642

Abstract

The rapid growth of e-commerce has led to a surge in user-generated product reviews, making manual sentiment analysis impractical. This study explores automated sentiment classification using two machine learning algorithms: Naïve Bayes and Support Vector Machine (SVM) that is optimized with Sequential Minimal Optimization (SMO). The dataset comprises 2,000 Shopee product reviews that are labeled as positive, neutral, or negative. The study focuses on assessing the effectiveness of these algorithms in classifying product reviews, especially in the diverse and high-volume data that is typically on e-commerce environments. Empirical evaluation shows that Naïve Bayes achieves 68% accuracy, while SVM with SMO attains 79%. Additionally, the study evaluates other important performance metrics, such as precision, recall, and F1-score. This study show that SVM with SMO outperforms Naïve Bayes in accurately classifying product reviews. These findings highlight the superior capability of SVM with SMO in handling complex sentiment data, thereby offering a more robust foundation for automated review classification. This research provides insights into selecting suitable classifiers for improving customer experience and strategic decision-making in digital commerce.
Residual-Based Hybrid SARIMA–LSTM for Bali Tourism Demand Forecasting Using Google Trends Junaedi; Aditiya Hermawan; Yusuf Kurnia; Ardiane Rossi Kurniawan Maranto
Journal of Information System and Informatics Vol 8 No 3 (2026): June
Publisher : Asosiasi Doktor Sistem Informasi Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.63158/journalisi.v8i3.1644

Abstract

Accurate tourism demand forecasting is essential for destinations characterized by strong seasonality, nonlinear fluctuations, and post-pandemic recovery uncertainty. This study develops a residual-based hybrid SARIMA–LSTM model for forecasting monthly international tourist arrivals to Bali, Indonesia, using historical arrival data and Google Trends search query data. The dataset covers January 2009 to December 2024, comprising 192 monthly observations. A chronological split was applied, with January 2009 to December 2022 used for training and January 2023 to December 2024 used for testing. SARIMA was employed to capture linear and seasonal structures, while LSTM was used to learn nonlinear residual patterns. The proposed model was compared with SARIMA, Random Forest, standalone LSTM, and SARIMA–RF using RMSE, MAPE, and R². The SARIMA–LSTM model achieved the best performance, with RMSE = 35,915.36, MAPE = 5.64%, and R² = 0.68, compared with SARIMA, which obtained RMSE = 37,052.68, MAPE = 5.70%, and R² = 0.65. These findings indicate that residual-based hybridisation provides incremental forecasting improvement. However, the independent contribution of Google Trends is not separately isolated in this study and should therefore be interpreted cautiously as a complementary behavioural signal within the proposed forecasting framework.
Enhancing digital asset ownership through decentralized non fungible token applications Yusuf Kurnia; Rino Rino; Edy Edy; Junaedi Junaedi; Aditiya Hermawan; Kevin Kevin
IAES International Journal of Artificial Intelligence (IJ-AI) Vol 15, No 2: April 2026
Publisher : Institute of Advanced Engineering and Science

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.11591/ijai.v15.i2.pp1972-1981

Abstract

The rapid expansion of the digital ecosystem has introduced pressing challenges surrounding identity, authenticity, trust, and transparency. The ease with which digital content can be duplicated often undermines creators, whose works are distributed without consent or fair compensation. Blockchain technology offers a transformative solution through its decentralized, transparent, and tamper-resistant structure. Among its innovations, non-fungible tokens (NFTs) provide a mechanism to verify the authenticity and ownership of unique digital assets. This study explores the transformative potential of NFTs in strengthening digital ownership and authenticity while identifying critical challenges such as market concentration, interoperability limitations, and security vulnerabilities within public NFT platforms. Employing the extreme programming (XP) methodology, this research proposes a secure framework for NFT creation outside public marketplaces to enhance the protection of smart contracts and user accounts. The findings demonstrate that this approach grants users’ greater control, minimizes exposure to platform-level risks, and promotes trust in decentralized asset management. Overall, this study underscores NFTs’ pivotal role in reshaping digital ownership models and highlights the need for continued innovation to ensure security, transparency, and equitable value distribution in the evolving digital economy.
Multimodal Wearable-Based Stress Detection Using Machine Learning: A Systematic Review of Validation Protocols and Generalization Gaps (2021 – 2025) Pannavira; Aditiya Hermawan
ULTIMATICS Vol 17 No 2 (2025): Ultimatics : Jurnal Teknik Informatika
Publisher : Faculty of Engineering and Informatics, Universitas Multimedia Nusantara

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.31937/ti.v17i2.4488

Abstract

Stress is a major determinant of mental health and productivity. Consequently, continuous, unobtrusive stress detection using wearable sensors and machine learning (ML) has become a key priority in digital health. This paper presents a Systematic Literature Review (SLR) of 19 peer-reviewed articles, selected from 36 initial papers via structured inclusion/exclusion criteria focusing on studies from 2021-2025 that report quantitative ML performance. We employed a quantitative and qualitative synthesis to analyze and map five key dimensions: sensing modalities, ML/DL algorithms, datasets, validation protocols, and societal feasibility. Findings reveal a clear state-of-the-art: multimodal physiological fusion (notably PPG, EDA, and ACC) paired with hybrid deep models (CNN-LSTM) consistently achieves the highest accuracy (85–96%) on benchmark datasets. Our research reveals a significant lab-to-field gap. Most studies utilize intra-subject or k-fold cross-validation, whereas the more robust Leave-One-Subject-Out (LOSO) validation is hardly employed, constraining model applicability. Furthermore, fewer than 15% of studies explicitly address vital practical constraints such as privacy, computational efficiency (Edge AI), or power consumption. This review methodically quantifies the gap, emphasizing that current models, despite their accuracy, are not yet suitable for real-world implementation. We conclude with actionable directions toward generalizable, lightweight, and privacy-aware stress-aware systems.
Co-Authors Ardiane Rossi Kurniawan Maranto Edy Edy Junaedi Junaedi Junaedi Kevin Kevin Pannavira Riki Rino Rino Sonya Eliesse Dameria Yusuf Kurnia