cover
Article Per Year (5 Year)
All
Home Page
OAI Link
Editorial Team
Contact
Reviewer
Google Scholar
Contact Name
Jumanto
Contact Email
jumanto@mail.unnes.ac.id
Phone
+6281339762820
Journal Mail Official
shmpublisher@gmail.com
Editorial Address
Jl. Karanglo Raya No. 64, Pedurungan, Semarang, 50191, Indonesia
Location
Kota semarang,
Jawa tengah
INDONESIA
Journal of Soft Computing Exploration
Published by shm publisher
ISSN : 27467686     EISSN : 27460991     DOI : https://doi.org/10.52465/joscex
Core Subject : Science, Engineering,
Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Electrical & Electronics Engineering
The journal focuses on publishing high-quality, original research and review articles in the field of Soft Computing, Informatics and Computer Science, emphasizing the development, application, and rigorous evaluation of Advanced Computational Methods, Artificial Intelligence (AI), Machine Learning (ML), and Data Science to address complex real-world challenges. The scope of the journal includes, but is not limited to, innovative research in the following areas: 1. Artificial Intelligence and Machine Learning Novel Algorithms and Architectures: Development and comparison of ML/DL models for classification and prediction (including Logistic Regression, Ridge Classifier, SVM, k-NN, and Random Forest). Ensemble Learning: Evaluation and optimization of ensemble methods Balanced Random Forest, SMOTE-RF, SMOTEBoost, and RUSBoost for robust prediction. Data Challenges and Preprocessing: Techniques for mitigating issues like class imbalance (using methods like SMOTE and GAN) and feature extraction/dimension reduction techniques (including Principal Component Analysis (PCA) and Local Binary Pattern (LBP)). 2. Deep Learning and Computer Vision Convolutional Neural Networks (CNNs): Research on CNN architectures (VGG16, ResNet50, DenseNet121, EfficientNet, and MobileNetV2) and the impact of optimization functions (Adam, SGD, NAdam) on model performance. Hybrid and Concatenated Architectures: Proposing and evaluating hybrid models (MobileNetV2 combined with LBP) or concatenated architectures (MobileNetV2 and DenseNet201) to improve classification and feature representation. Image Analysis Tasks: Advanced techniques for image classification (specifically Diabetic Retinopathy), image similarity detection (using Siamese Networks and Test-Time Augmentation), and multi-object segmentation (using FCN with Squeeze-and-Excitation and Attention Mechanisms for palm oil images). 3. Data Science and Advanced Analytics Pattern Detection and Data Mining: Performance evaluation of data mining algorithms, including Biclustering (Cheng & Church and Spectral Biclustering), specifically under challenging structural conditions like collinearity and overlap. Time Series Analysis and Forecasting: Application of advanced decomposition and clustering methods (Ensemble Empirical Mode Decomposition (EEMD) and Time Series Clustering with DTW/ARIMA) for accurate economic or temporal prediction. 4. Applied Informatics (Domain-Specific Applications) Health and Medical Informatics: Classification models for disease diagnosis (including Heart Attack Disease and Diabetic Retinopathy). Agricultural Informatics: Automated detection and classification of plant diseases from leaf/crop images (including Mango Leaf Disease and Chili Plant Disease) and Palm Oil Segmentation. Business and Economic Informatics: Predictive modeling for crucial business metrics (Customer Churn Prediction in Telecommunications) and economic forecasting (Rice Price Forecasting).
Arjuna Subject : Ilmu Komputer (semua kategori) - Kecerdasan Buatan
Articles 74 Documents
 4   5   6   7   8 
A hybrid VGG16 and random forest model for multi-class ischemic heart disease detection via ecg image analysis Rizka Dian Safitri; Christy Atika Sari; Noorayisahbe Mohd Yaacob
Journal of Soft Computing Exploration Vol. 7 No. 3 (2026): September 2026
Publisher : SHM Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52465/joscex.v7i3.121

Abstract

A cardiac illnesses, mainly ischemic cardiac conditions, are a primary factor behind global fatality rates. Clinical diagnosis generally relies on manual interpretation of Electrocardiogram (ECG) signals, which is subjective and time-consuming, compounded by the challenge of limited access to raw ECG data on commercial devices. This research focuses on designing an automated diagnosis system based on ECG image analysis using the VGG16 Convolutional Neural Network (CNN) architecture through a knowledge transfer method. A total of 8,268 ECG images from 689 patients, divided into three categories Normal, Abnormal Heartbeat, and History of Myocardial Infarction (MI) were evaluated in this study. Test results demonstrated that the VGG16 architecture integrated with Random Forest produced the most optimal performance, with a test accuracy of 93.27%, an F1-Score of 93.25%, along with an Area Under the Curve (AUC) value of 0.991. This combined model successfully detected Normal images without any prediction errors. This computational image feature extraction approach has proven effective in reducing diagnostic subjectivity and holds strong potential for application as a fast and consistent clinical decision support system across various healthcare facilities. Nevertheless, multicenter validation on a more diverse population is still required to ensure the model's clinical generalizability.
Rethinking the role of three-star ratings: Handling inconsistency in indonesian tourism reviews using cost-sensitive learning XGBoost Muhammad Nur Hikmah Ramadhan; Akhmad Syaifuddin; Ristu Saptono
Journal of Soft Computing Exploration Vol. 7 No. 3 (2026): September 2026
Publisher : SHM Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52465/joscex.v7i3.159

Abstract

Indonesian people often give contrary online reviews, for instance ratings that do not always match the actual feelings. This can make it difficult for tourism sector, such as Solo Safari, to handle complaints and improve service quality. Consequently, this study has formulated a rating prediction model using the XGBoost algorithm with a dataset of 2,047 reviews that have been relabeled. The best model, the FastText-XGBoost with Cost-Sensitive Learning, signify that it performs quite well with an average difference the prediction is only 0.04 points from the actual rating. However, the results are not optimal because the meaning of the review text still feels fuzzy even though being relabeled. This problem arises because reviewers tend to position sentiment very positively or negatively in the moderate category, so the perimeters between classes become less clear. This study then proposed an extreme class restructuring by simplifying the category by removing the 3-star rating. This method can increase the accuracy of the model to 0.9440 and clarify the category limits. Therefore, the model can be used on the service dashboard to help Solo Safari management respond to critical feedback faster.
Metaheuristic optimization in AHP–TOPSIS pairwise matrices for decision support on preventing low birth weight in adolescent pregnancies Muhammad Annys; Wiharto; Shaifudin Zuhdi
Journal of Soft Computing Exploration Vol. 7 No. 3 (2026): September 2026
Publisher : SHM Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52465/joscex.v7i3.160

Abstract

Low Birth Weight (LBW), defined as a birth weight below 2,500 grams, remains a major public health concern in developing countries due to its association with increased neonatal morbidity and mortality. Decision-support systems based on the Analytic Hierarchy Process (AHP) and Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) have been widely applied in healthcare prioritization; however, their reliability may be compromised by inconsistencies in expert-generated pairwise comparison matrices. Despite the growing use of metaheuristic optimization techniques, limited studies have compared the effectiveness of Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) in improving AHP–TOPSIS weighting consistency for LBW prevention. This study aims to compare the performance of GA and PSO for optimizing pairwise comparison matrices within a hybrid AHP–TOPSIS framework for adolescent-pregnancy intervention prioritization. The study utilized maternal and child health data obtained from the Indramayu District Health Office in 2025. Criterion weights were determined using AHP, while intervention alternatives were ranked using TOPSIS. GA and PSO were applied to minimize matrix inconsistency through hyperparameter-optimized search processes. Performance was evaluated using Consistency Ratio (CR), Hamming Distance, Euclidean Distance, Kendall’s Tau, and computational time. The results showed that both GA and PSO successfully improved matrix consistency and generated alternative feature-weight distributions while maintaining moderate agreement with the baseline ranking structure. GA achieved higher ranking stability (Kendall’s Tau = 0.7197) and lower computational time, whereas PSO produced greater weight redistribution. These findings demonstrate that metaheuristic optimization can enhance the robustness and consistency of AHP–TOPSIS-based weighting schemes, providing a more reliable ranking-based decision-support mechanism for LBW prevention and intervention prioritization.
Model bidirectional GRU with bayesian optimization for dry gas pressure forecasting in transmission pipeline networks Mohammad Alvinanda Kurniawan; Aditya Firman Ihsan; I Wayan Palton Anuwiksa
Journal of Soft Computing Exploration Vol. 7 No. 3 (2026): September 2026
Publisher : SHM Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52465/joscex.v7i3.166

Abstract

Operational robustness and infrastructure safety require that stable dry gas pressure in transmission pipeline networks be maintained. Due to the highly nonlinear and time-varying behaviors of the dry gas pressure, conventional methods cannot predict its tendencies accurately. Deep learning models have recently gained good performance for time-series forecasting, however combination of Bidirectional Gated Recurrent Unit (BiGRU) and Bayesian Optimization based approach for dry gas pressure forecasting remains underexplored. Based on these, this study develops Bayesian optimization-enhanced forecasting framework for BiGRU to enhance the prediction performance. The framework is based on a Gaussian Process surrogate model and an acquisition function — in this case Expected Improvement, which guides the search for optimal hyperparameter configurations. We used operational SCADA time-series data consisting of hourly measurements (pressure, temperature, flowrate and gas composition) collected from a dry gas transmission pipeline across two annual periods. Pre-processing of data consists of dealing with missing data, Min–Max normalization and sliding window conversion. Here a multistep forecasting scenario for the next 20 hours was used, and thismultistep prediction was evaluated calculating the MAE, RMSE, and R². The optimized BiGRU performed the MAE of 11.7729 psi, RMSE of 14.7827 psi and R² of 0.9389, which enhanced the baseline model by 13.67%, 12.90%, and 1.60 percentage points respectively. These results indicate that Bayesian Optimization improves the forecasting performance of BiGRU and, at the same time, decreases the manual hyperparameter tuning efforts.
Security assessment and security hardening of a laravel-based e-government web application using OWASP top 10 Muh Firdaus Garra Daeng Sikki; Adzanil Rachmadhi Putra; Rizqi Ramadhan
Journal of Soft Computing Exploration Vol. 7 No. 3 (2026): September 2026
Publisher : SHM Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52465/joscex.v7i3.210

Abstract

The rapid adoption of e-government web applications has increased the need for effective cybersecurity measures to protect public services and sensitive information from evolving cyber threats. This study aims to identify security vulnerabilities, evaluate their severity, and implement security hardening measures for a Laravel-based e-government web application using the OWASP Top 10 (2021) framework. A Vulnerability Assessment and Penetration Testing (VAPT) approach was applied by combining automated vulnerability scanning with White Box and Grey Box penetration testing. Identified vulnerabilities were classified using the OWASP Top 10 (2021), prioritized with the Common Vulnerability Scoring System (CVSS) v3.1, and addressed through targeted security hardening followed by security verification. The assessment identified six High-severity vulnerabilities (CVSS v3.1 scores 7.1–8.8) across six OWASP categories. The implemented security controls, including HTTPS enforcement, input sanitization, security headers, rate limiting, dependency updates, and stronger authentication, successfully mitigated all identified vulnerabilities and improved the application's overall security posture. Overall, integrating OWASP-based security assessment with systematic security hardening provides an effective and practical strategy for improving the security resilience of Laravel-based e-government web applications.
Engagement–sentiment gap analysis of tiktok skincare content using IndoBERT: a comparative study of random forest and XGBoost Dwita Amalia Rizki; Andy Prasetyo Utomo; Zainur Romadhon
Journal of Soft Computing Exploration Vol. 7 No. 3 (2026): September 2026
Publisher : SHM Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52465/joscex.v7i3.212

Abstract

Social media platforms, particularly TikTok, have become important sources of consumer-generated information that influence product awareness and purchasing decisions in the skincare industry. However, audience engagement metrics alone cannot fully explain audience perception because highly engaging content does not always receive positive responses from users. This study aims to propose and evaluate an Engagement–Sentiment Gap framework that integrates audience interaction and audience perception to provide a more comprehensive evaluation of TikTok skincare content than conventional sentiment analysis or engagement-based approaches. A dataset comprising 20,688 comments and metadata from 603 TikTok videos across ten skincare-related topics was collected through web scraping. After text preprocessing, sentiment classification was performed using a fine-tuned IndoBERT model, followed by video-level sentiment aggregation, engagement score calculation, Engagement–Sentiment Gap categorization, and engagement-level prediction using Random Forest and XGBoost. The results showed that neutral sentiment dominated the discussions (72.5%), while audience sentiment exhibited a statistically significant but weak positive correlation with engagement (r = 0.2283, p < 0.001). Random Forest achieved the best predictive performance, with 65.27% Balanced Accuracy and 68.74% ROC-AUC, while topic was identified as the most influential predictive feature. These findings demonstrate that integrating sentiment and engagement within the proposed framework provides richer insights into audience behavior and content performance than evaluating engagement or sentiment independently, offering a practical analytical approach for social media content evaluation.
Performance evaluation of DETR and YOLO models in multilabel detection of motorcycles and license plates across varying resolutions Faisal Fajri Rahani; Miftahurrahma Rosyda
Journal of Soft Computing Exploration Vol. 7 No. 3 (2026): September 2026
Publisher : SHM Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52465/joscex.v7i3.75

Abstract

Two-wheeled vehicles dominated the transportation landscape in Indonesia, yet automated monitoring faced challenges due to the complexity of multilabel detection in dense traffic conditions. This research aimed to conduct a comparative analysis between the Detection Transformer (DETR) and You Only Look Once (YOLO) models for multilabel object detection on motorcycles, specifically identifying vehicle types and license plates simultaneously. The research workflow involved multi-label annotating 550 surveillance images, partitioning the dataset using a 70:20:10 hold-out ratio, and applying spatial augmentations (zooming and shearing) to enhance generalization. Subsequently, the DETR, YOLOv11, and YOLOv12 architectures were trained and evaluated across six input resolution scales ranging from 128px to 640px based on mean Average Precision (mAP) metrics and training convergence rates. The results indicated that the DETR model at 640px resolution achieved the highest performance with a mean Average Precision (mAP@50) of 96.97%. Another significant finding revealed that DETR exhibited superior training efficiency compared to the YOLO family, where optimal mAP convergence was reached within only 1 to 7 epochs. In contrast, YOLO models required significantly more training iterations to achieve comparable accuracy levels. This study concluded that the attention mechanism in transformer-based architectures provided higher effectiveness in learning multilabel features for two-wheeled vehicles than conventional convolutional approaches.
Sentiment analysis and topic modeling of halal traveler satisfaction factors: a computational approach based on latent dirichlet allocation (LDA) and naïve bayes Ahmad Ashril Rizal; Siti Rabi&#039;atul Adawiyah
Journal of Soft Computing Exploration Vol. 7 No. 3 (2026): September 2026
Publisher : SHM Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52465/joscex.v7i3.95

Abstract

The rapid growth of global halal tourism positions West Nusa Tenggara (NTB) Province strategically, requiring a rapid, adaptive system for evaluating service quality to meet tourists' digital needs. However, local government evaluations are hampered by the limitations of conventional sentiment analysis methods, which often generalise textual emotions without specifically examining the dimensions of sharia services. Therefore, this study develops an integrated hybrid model combining topic modelling algorithms and sentiment analysis for the context of halal destinations in NTB. This R&D study integrates Selenium-based text mining and web scraping to extract unstructured social media and online news data, Latent Dirichlet Allocation (LDA) to identify automated factors, and the Naïve Bayes algorithm to calculate the sentiment polarity of each factor. The LDA model successfully identified five key factors of tourist satisfaction. The sentiment analysis results show that digital responses are dominated by positive (38.0%) and neutral (61.0%) sentiment, with the highest satisfaction being security (71.4%), while infrastructure and health are critical areas for improvement. The Interactive Monitoring Dashboard, the output of this research, has been validated for usability using the System Usability Scale (SUS) with 15 NTB tourism stakeholders, achieving a SUS score of 88.0 and entering the "Very Good" (Grade A) category.
HCViT: A hybrid convolutional vision transformer with spatial attention bias for brain tumor classification in MRI Erfan Wahyudi; Bahtiar Imran
Journal of Soft Computing Exploration Vol. 7 No. 3 (2026): September 2026
Publisher : SHM Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52465/joscex.v7i3.104

Abstract

Although deep learning has significantly advanced MRI-based brain tumor classification, computational feasibility for clinical deployment remains bottlenecked by the heavy parameter cost of Vision Transformers and the limited global context awareness of Convolutional Neural Networks. We propose HCViT (Hybrid Convolutional Vision Transformer), a novel architecture designed for medical imaging by integrating local feature extraction (ConvStem), global context modeling (Transformer), and spatially-aware attention via learnable convolutional bias (ConvBias). This domain-aware hybridization preserves anatomical details while capturing long-range dependencies, overcoming key limitations of pure CNNs and Transformers. Evaluated on binary brain MRI classification (tumor vs. no tumor), HCViT achieves 96.25% validation accuracy, matching Hybrid ViT (ResNet26d + ViT), the top baseline, while being 7× more parameter-efficient (8.99M vs 62.78M) and 2.5× more computationally efficient (2.012 vs 5.049 GFLOPs). It surpasses ResNet-50 (93.75%), ConvNeXt-Tiny (92.50%), ViT-Tiny (83.75%), ViT-Base (68.75%), and Swin-T (78.75%). With 2032 img/s inference throughput (batch=32), HCViT offers clinically viable speed without compromising diagnostic precision. Its robust generalization and high sensitivity to subtle lesions make it suitable for real-world deployment, including resource-limited or high-throughput clinical settings. This work demonstrates that domain-informed architectural design, not merely adopting trending models, is essential for effective, efficient, and interpretable medical AI. HCViT establishes a new state-of-the-art efficiency-accuracy trade-off for brain tumor classification and provides a scalable foundation for future extensions: multi-class diagnosis, integration with segmentation models, and edge-device deployment. By bridging CNNs’ local inductive bias with Transformers’ global modeling, enhanced by spatial attention, HCViT sets a benchmark for next-generation medical vision models.
OCR-integrated retrieval-augmented generation for PDF-document question answering system in vocational higher education Mutiara Auliya Khadija; Muhamad Prayoga Putra Setyawandana; Muhammad Raihan; Wahyu Nurharjadmo; Ina Primasari Primasari; Abdul Aziz Aziz
Journal of Soft Computing Exploration Vol. 7 No. 3 (2026): September 2026
Publisher : SHM Publisher

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52465/joscex.v7i3.225

Abstract

Institutional documents in vocational higher education are generally stored in PDF format with heterogeneous characteristics, including both digital and scanned documents, making the manual information retrieval process less efficient. This study aims to develop an OCR-Integrated Retrieval-Augmented Generation (RAG) system to support PDF-document question answering that can provide accurate and traceable document-based answers. The proposed framework consists of three main stages such as document ingestion and processing, retrieval-augmented question answering, and evaluation. The system integrates text extraction using PyMuPDF, Google Tesseract (v5.3.3) as a OCR fallback for the documents, text cleaning, chunking, embedding formation, vector storage with Facebook AI Similarity Search (FAISS), semantic retrieval, prompt augmentation, and answer generation. Evaluation is carried out using Retrieval Augmented Generation Assessment (RAGAS) Evaluation, BERTScore F1 and Bilingual Evaluation Understudy (BLEU) Score. This experiment using 50 total of PDF-documents related with various institutional documents from Vocational Higher Education with 150 test question-answer queries. The results show that the Multilingual-e5-base embedding model has best performed with a context precision of 0.8333, a faithfulness score of 1.0000 and the BERTScore F1 score reached 0.6494.  These research indicate that OCR-Integration plays a critical role in enabling the RAG pipeline to improves retrieval and answer-generation performance.

Page 6 of 8 | Total Record : 74

 4   5   6   7   8 

Filter by Year

2026 2026


Reset
Filter By Issues
All Issue Vol. 7 No. 4 (2026): December 2026 Vol. 7 No. 3 (2026): September 2026 Vol. 7 No. 2 (2026): June 2026 Vol. 7 No. 1 (2026): March 2026