Articles

Access the latest knowledge in applied science, electrical engineering, computer science and information technology, education, and health.

Advanced Search
Filter Icon

Filters article

Years

FAQ Arrow
Single Year
Range Year
0
0

Source Title

FAQ Arrow

Authors

FAQ Arrow

29,734 Article Results

Level of detail in UML models and its impact on model comprehension: a replication study

10.11591/ijeecs.v41.i3.pp1095-1104
Ariadi Nugroho , Michel R.V Chaudron
This replication study examines the impact of level of detail (LoD) in unified modeling language (UML) on model comprehension, replicating a controlled experiment, which involved 53 MSc students at Eindhoven University of Technology. Using the same UML model and experimental design, we conducted the study with 23 MSc Computer Science students at Bina Nusantara University, Indonesia. Consistent with the original findings, higher LoD was found to enhance comprehension correctness. However, the effect on comprehension efficiency was weaker and not statistically significant, likely due to the smaller sample size and contextual differences in subjects’ backgrounds. Furthermore, we found a potential disconnect between perception and actual comprehension performance in the subjects receiving UML model with low LoD. Specifically, while they viewed the model favourably, their actual understanding may have been impaired by the limited information and therefore the perceived clarity and ease of comprehension are not reflective of the true comprehension. Overall, this study reinforces the importance of LoD in UML modeling and highlights the need for further replication, particularly in contexts involving professional software engineers.
Volume: 41
Issue: 3
Page: 1095-1104
Publish at: 2026-03-10

A hybrid approach for measuring semantic similarity in lexically identical but ambiguous sentences

10.11591/ijeecs.v41.i3.pp954-965
Btissam El Janati , Adil Enaanai , Fadoua Ghanimi
This study addresses the critical challenge of semantic similarity and lexical disambiguation in natural language processing, focusing on sentences with structural and lexical ambiguities. We introduce an innovative hybrid approach that synergistically combines symbolic and neural methods to better align with human judgment. Our methodology dynamically integrates fuzzy Jaccard’s lexical precision with SBERT embeddings’ contextual sensitivity, enabling adaptive semantic ambiguity resolution. Experimental evaluation on 33 ambiguous sentences demonstrates that our approach significantly outperforms conventional artificial intelligence (AI) systems, achieving an 11.7% reduction in mean absolute error compared to reference models, with statistical analysis confirming robust results (d = -0.80, p < 0.001). This represents a 65% improvement in human evaluation alignment over existing methods. Our research contributes to advancing the field by showing that architectural intelligence can surpass mere parameter scaling, offering an effective solution for applications requiring both precision and interpretability, with promising directions for multilingual extension and explainable AI integration.
Volume: 41
Issue: 3
Page: 954-965
Publish at: 2026-03-10

An innovative deep learning based approach for anomaly detection in intelligent video surveillance

10.11591/ijeecs.v41.i3.pp1105-1116
Megha G. Pallewar , Vijaya R. Pawar
Nowadays, anomaly detection has gained vital importance as security is a major concern everywhere. This work focuses on developing an intelligent video surveillance system capable of detecting anomalous activities in videos, utilizing the UCF Crime dataset as the primary source. The proposed model employed a multistage method uniting the convolutional neural networks (CNN) and long short-term memory (LSTM) networks. In the proposed approach, video frames serve as input to the CNN, which processes them to extract key features. These features are then passed to an LSTM network to capture temporal dependencies and identify anomalous events over time. This CNN-LSTM architecture successfully detects twelve distinct types of anomalous activities: abuse, arrest, arson, assault, burglary, explosion, fight, road accident, robbery, stealing, shoplifting, and vandalism. The dataset is divided into portions for training, testing, and validation, along with cross-validation to ensure model generalization. The system achieves an accuracy of 98.6%, reflecting a significant improvement of 4-5% over existing systems. This demonstrates the robustness of the proposed method in detecting anomalous behavior in video data.
Volume: 41
Issue: 3
Page: 1105-1116
Publish at: 2026-03-10

Intelligent artificial neural network-based control for solar electric vehicle charger

10.11591/ijeecs.v41.i3.pp885-893
Rajeshkumar Damodharan , Pradeep Kumar S
The performance of electric vehicle (EV) charging systems in response to sudden changes in solar irradiation and dynamic battery load variations. EV chargers must have effective power conversion and flexibility as the use of renewable energy sources increases. This paper suggests a charging system based on resonant converters that minimizes heat and losses in EV charging stations by enabling high-efficiency, soft-switching power transfer. For modern EV applications, the ability to manage large voltage fluctuations ensures reliable, quick, and portable charging. The artificial neural networks (ANN) controller overcomes the drawbacks of conventional Perturb and Observe (P&O) for solar DC-DC converters and PI control for resonant converter approaches. MATLAB simulation results demonstrate that the proposed system outperforms traditional techniques in terms of an ANN based controller, which enhances maximum power point tracking (MPPT) efficiency to 98.6%, reduces oscillations near the maximum power point by approximately 80%, and increases total EV charging efficiency by 3%. The ANN-based control to EV charging infrastructure greatly enhances overall system dependability and real-time responsiveness, making it a good fit for subsequent smart grid and renewable energy applications.
Volume: 41
Issue: 3
Page: 885-893
Publish at: 2026-03-10

Exploring word embeddings and clustering algorithms for user reviews

10.11591/ijeecs.v41.i3.pp1017-1024
Zuleaizal Sidek , Sharifah Sakinah Syed Ahmad
The rapid advancement of information technology has led to a significant surge in the volume of unstructured textual data. This has posed a major problem in terms of analyzing, organizing, and automatically clustering text for research purposes, which is crucial for extracting valuable insights. The process of manually clustering the unstructured data, such as customer reviews on the Internet, which capture the opinions of customers regarding products, services, and social events, requires significant financial resources, manpower, and time. Most of the studies are directed towards the analysis of sentiment in user reviews. In order to address the issues effectively, automated text clustering could assist in categorizing reviews into various themes, thereby simplifying the analysis process. Therefore, in this paper, we present and compare the result of experiment the combination of five text clustering techniques, namely K-means, fuzzy C-mean (FCM), non-negative matrix factorization (NMF), latent dirichlet allocation (LDA), and latent semantic analysis (LSA) with different embedding techniques, namely term frequency–inverse document frequency (TF-IDF), Word2Vec, and global vectors (GloVe). The experiments revealed that LDA is a reliable algorithm as it consistently produces good results across three-word embeddings. The highest Silhouette score recorded in the experiments was 0.66 using LDA and Word2Vec as word embedding. Simultaneously, the application of LSA in conjunction with Word2Vec yields superior outcomes, as evidenced by a Silhouette score of 0.65.
Volume: 41
Issue: 3
Page: 1017-1024
Publish at: 2026-03-10

Satellite-based assisted-offloading for energy-constrained edge networks

10.11591/ijeecs.v41.i3.pp935-945
Thembelihle Dlamini , Mengistu A. Mulatu , Sifiso Vilakati
As the need for global broadband internet connectivity increases, there is a need to consider the use of non-terrestrial networks (NTNs) to extend the network coverage to protected areas (e.g., national parks). Usually, protected areas are prohibited from having power lines thus lacking wireless connectivity. To over come this challenge, energy can be provided through the use of green energy from a solar photovoltaic (PV) system. Then, a green energy-based base station (BS) can be deployed within the area in order to provide mobile connectivity to visitors, as well as also using the NTNs to handle excess traffic or take over the traffic in the event the BS does not have sufficient green energy from stor age. In this paper, a hybrid wireless communication system is proposed to in clude BS sites located in a protected area and satellites in the low earth orbits (LEO), coupled with new offloading strategies, with the main goal of optimizing the trade-off between energy consumption and end-to-end delay for the green energy-based BS sites. For accuracy of our simulations, we consider real data from a solar photovoltaics system, traffic workloads, visitor’s location data, and satellite orbits from Starlink constellations. Our results demonstrate that the co existence of the BS and satellite achieve energy savings from 59% to 34%, with an average system delay of 0.83 seconds and a packet drop rate that ranges from 8.3% to 2.7%, when compared with our benchmark.
Volume: 41
Issue: 3
Page: 935-945
Publish at: 2026-03-10

Synthetic inertia controller of a wind power plant as a means of increasing the stability of electric power systems

10.11591/ijeecs.v41.i3.pp1117-1123
Makhmudov Tokhir Farkhadovich , Ramatov Adxam Nasiriddin o’gli
The article discusses the use of wind power plants as sources of synthetic inertia to enhance power system stability and reduce frequency fluctuations. This research explores the feasibility of implementing a synthetic inertia controller in wind power plants to decrease the magnitude of frequency oscillations during transient operating conditions. The growing integration of wind farms into modern power grids leads to a reduction in the overall kinetic energy, or inertia, available in the system. As a result, the grid may become more vulnerable to disturbances. When the system inertia is too low, frequency stability can be affected, especially when large generating units suddenly fail or disconnect from the grid. In general, a lower level of inertia in the system causes larger frequency deviations following an imbalance in active power. To overcome this issue, a synthetic inertia regulator for wind power plants has been developed, enabling wind turbines to support the grid and reduce the depth of frequency drops during transient events.
Volume: 41
Issue: 3
Page: 1117-1123
Publish at: 2026-03-10

Fuzzy logic-based load balancing for voltage symmetry in distribution networks

10.11591/ijeecs.v41.i3.pp873-884
Adeel Saleem , Kholiddinov Ilkhombek Khosiljonovich , Kholiddinova Mashkhurakhon Mutalibjon Qizi , Begmatova Mukhlisakhon Mutalibjon Qizi , Sharobiddinov Mirzokhid
This paper introduces a load balancing approach based on fuzzy logic to enhance the efficiency of power distribution networks. The unbalance of voltages and an unequal load of the phases continue to be the problematic situation of the low-voltage distribution networks, particularly as the percentage of photovoltaic (PV) systems is growing. The results of such conditions include a deviation of voltage, higher losses of power, faster equipment aging, and lower power quality. This paper proposes a fuzzy logic–based phase load balancing approach that explicitly integrates voltage symmetry requirements defined by the GOST 13109-97 power quality standard. Unlike optimization-based and heuristic methods, the proposed fuzzy logic controller (FLC) redistributes phase currents using linguistic rules derived from voltage unbalance coefficients and phase current conditions, without iterative optimization procedures. Simulation results obtained in MATLAB/Simulink demonstrate a reduction of the voltage unbalance factor (VUF) by approximately 25–30% and a decrease in active power losses by 12–15% compared to the initial unbalanced operating state. The proposed method offers low computational complexity, fast response, and high interpret-ability, making it suitable for real-time implementation in smart distribution networks with distributed PV generation.
Volume: 41
Issue: 3
Page: 873-884
Publish at: 2026-03-10

A multimodal framework for explainable chest X-ray report generation

10.11591/ijeecs.v41.i3.pp1060-1069
Hamza Chehili , Nourhene Bougourzi , raida malak Makhlouf , hadjer Taib , Mustapha Bensaada
Chest X-ray (CXR) interpretation remains a challenging task due to overlapping anatomical structures, variability in disease presentation, and increasing clinical workload. Existing automated report-generation models provide promising results but often lack explicit interpretability, limited clinical alignment, and insufficient comparative evaluation with established baselines. This study proposes an explainable multimodal framework that combines a dual CNN encoder (ResNet-50 and EfficientNet-B0) with the Gemma-3 1B language model fine-tuned using low-rank adaptation (LoRA). Visual explanations are produced through Gradient-weighted Class Activation Mapping (Grad-CAM) to enhance transparency in the decision process. Unlike prior image-to-text pipelines, our approach follows a findings-guided paradigm and integrates both visual and textual cues during generation. Experiments conducted on public datasets demonstrate consistent improvements over representative vision-language baselines reported in recent literature, with notable gains in BLEU, ROUGE, METEOR, and BERTScore. Generated reports show improved factual completeness and clinically relevant region-level attention. Limitations include the absence of evaluation against emerging foundation models and the need for anatomical- level explainability metrics. Future work will extend benchmarking to models such as M2-Transformer, MedCLIP-GPT, and R2Gen, and will explore clinical validation in real-world workflows.
Volume: 41
Issue: 3
Page: 1060-1069
Publish at: 2026-03-10

ARX based cipher with S-box augmentation: statistical and differential evaluation

10.11591/ijeecs.v41.i3.pp946-953
Manita Rajput , Pranali Chaudhari
With the growth of internet of medical things (IoMT), the continuous transfer of vital biomedical data requires lightweight encryption with strong resistance to statistical and differential attacks. The Speck cipher is a suitable candidate because of its low memory and execution time. However, its vulnerability to differential cryptanalysis limits wider use in healthcare environments. In this work, a hybrid lightweight algorithm is proposed by integrating the PRESENT substitution box within the Speck64/96 round structure. The substitution layer was evaluated at three different positions in the round function. Statistical and differential analyses were performed on four sets of plaintext data, each containing 1,000 test pairs. Index of coincidence (IoC), entropy, and avalanche effect were used as the primary statistical metrics. Differential trail strength was assessed using ciphertext differences and round-wise differential probability (DP). The experimental results show that the proposed version, named Speckpres_S, achieves a 6.02% reduction in IoC, a 3.8% improvement in entropy, and a 1.7% rise in avalanche effect when compared with Speck64/96. The differential trail becomes weaker, with a 46% reduction in trail probability and a 12–15% increase in trail weight across all datasets. The execution time remained within IoMT limits. This indicates stronger resistance to differential attacks with predictable diffusion. The study demonstrates that Speckpres_S improves security while maintaining practical latency and throughput for IoMT applications. Although execution time increases marginally, the gain in differential resistance and statistical performance makes the proposed algorithm a more robust option for transmitting sensitive biomedical parameters.
Volume: 41
Issue: 3
Page: 946-953
Publish at: 2026-03-10

Leveraging CNN to analyze facial expressions for academic engagement monitoring with insights from the multi-source academic affective engagement dataset

10.11591/ijeecs.v41.i3.pp977-999
Noora C. T. , P. Tamil Selvan
The dynamics of student engagement and emotional states significantly influence learning outcomes. Positive emotions, stemming from successful task completion, contrast with negative emotions arising from learning struggles or failures. Effective transitions to engagement occur upon problem resolution, while unresolved issues lead to frustration and subsequent boredom. Facial engagement monitoring is crucial for assessing students’ attention, interest, and emotional responses during learning. Recent advancements in convolutional neural networks (CNNs) show promise in automatically analyzing facial expressions to infer engagement levels. This study proposes a CNN-based approach utilizing the multi-source academic affective engagement dataset (MAAED) to categorize facial expressions into boredom, confusion, frustration, and yawning. By extracting features from facial images, this method offers an efficient and objective means to gauge student engagement. Recognizing and addressing negative affective states, such as confusion and boredom, is fundamental in creating supportive learning environments. Through automated frame extraction and model comparison, this study demonstrates reduced loss values with improving accuracy, showcasing the effectiveness of this method in objectively evaluating student engagement. Facial engagement monitoring with CNNs, using the MAAED dataset, is pivotal in understanding human behavior and enhancing educational experiences. The CNN model, trained on MAAED annotated facial expressions, accurately classifies engagement categories. Experimental results underscore the CNN-based approach’s efficacy in monitoring facial engagement, highlighting its potential to enrich educational environments and personalized learning experiences.
Volume: 41
Issue: 3
Page: 977-999
Publish at: 2026-03-10

Study of performance the 3-phase induction motor that drives by using scalar and vector control with variable speed loading

10.11591/ijeecs.v41.i3.pp894-911
Omran Alabedalkhamis , Baran Karahan , İbrahim İdiz , Hüseyin Alptekin , Enver Ediz Erol
Induction motor performance and efficiency greatly depend on the applied control technique, particularly in variable- and fixed-speed industrial applications. This paper aims to comparatively assess scalar control and vector control strategies for three-phase squirrel-cage induction motors. Using a simulation-based approach in MATLAB/Simulink, scalar control with sinusoidal pulse width modulation (SPWM) and vector control with space vector modulation (SVM) are built and analyzed under constant, variable, and bidirectional speed loading situations characteristic of a drive system. The results demonstrate that vector control provides greater speed regulation (about 93% compared to scalar control), reduced torque ripple (about 97% compared to scalar control), lower current stress (about 94% compared to scalar control), and improved dynamic responsiveness compared to scalar control, especially during transient operation. The paper is limited to numerical simulations. This paper’s biggest contribution is a clear, practical comparison which provides performance- and cost-oriented guidelines for selecting appropriate induction motor control strategies in severel applications.
Volume: 41
Issue: 3
Page: 894-911
Publish at: 2026-03-10

FGMPSO: a hybrid firefly-gradient-MOPSO framework for high-dimensional feature selection

10.11591/ijeecs.v41.i3.pp1082-1094
Alwatben Batoul Rashed
When working with high-dimensional datasets, selecting the most relevant features is essential for improving both model clarity and processing efficiency, all while keeping predictive accuracy intact. In response to this challenge, the study introduces firefly-gradient-multi-objective particle swarm optimization (FGMPSO), an advanced hybrid technique that blends the firefly algorithm, gradient descent (GD), and multi-objective particle swarm optimization (MOPSO). This approach is specifically designed to identify an optimal subset of features that balances dimensionality reduction with strong classification performance. The method was evaluated on eight benchmark datasets and compared against multiple PSO-based feature selection techniques. The empirical results demonstrated that FGMPSO consistently achieved superior or competitive classification accuracy while selecting significantly fewer features. Notably, in several datasets, FGMPSO not only reduced dimensionality but also outperformed other methods in terms of classification accuracy. This efficiency is attributed to the intelligent exploration of the search space by the firefly algorithm, refinement via GD, and effective trade-off optimization enabled by MOPSO. The findings suggest that FGMPSO is a robust and scalable solution for feature selection, particularly suitable for complex and high-dimensional datasets. Its adaptability, convergence speed, and balance between dimensionality reduction and accuracy position it as a valuable tool in modern machine learning pipelines.
Volume: 41
Issue: 3
Page: 1082-1094
Publish at: 2026-03-10

Behavioral analysis across multiple domains using machine learning and deep learning models

10.11591/ijeecs.v41.i3.pp1124-1133
Suryakant Suryakant , Kumar P K
Behavioral analysis.using machine learning (ML) and deep learning (DL) has become critical across healthcare, finance, cybersecurity, education, and marketing. This systematic review synthesizes advancements in ML- and DL-driven behavioral analysis (2019-2025) across five key domains. Our findings reveal that Deep Learning techniques achieve superior predictive accuracy (85-97% in healthcare imaging anomaly detection), while Machine Learning remains preferred for interpretability in finance (accuracy: 78-92%, with explainability advantage). A major trade-off emerges: DL models demonstrate higher accuracy but require substantial labeled data and computational resources, whereas ML models offer transparency but limited scalability. This review contributes by: (1) systematically analyzing domain-specific performance metrics and model evolution; (2) providing comparative synthesis of ML vs. DL approaches with quantitative benchmarking; (3) identifying critical challenges (data quality, privacy, algorithmic bias, interpretability); and (4) proposing actionable future directions, including Explainable AI, Federated Learning, and multimodal fusion. We adopt PRISMA-guided methodology examining 100+ peer-reviewed studies, revealing that hybrid ML-DL architectures represent the emerging best practice for balancing accuracy with interpretability.
Volume: 41
Issue: 3
Page: 1124-1133
Publish at: 2026-03-10

Ontology-based semantic link prediction for enhancing academic collaboration through knowledge management

10.11591/ijeecs.v41.i3.pp1040-1048
Pham Thi Thu Thuy , Thinh Thi Thuy
This paper introduces a novel ontology-based semantic link prediction framework that unifies structural, temporal, and semantic signals from heterogeneous scholarly sources to enhance academic collaboration forecasting. By integrating AMiner, DBLP, and Mendeley datasets into a unified SKOS- and Dublin Core-aligned ontology, the framework enables semantic enrichment, cross-source reasoning, and contextualized link prediction. Unlike previous studies that focus solely on structural features or basic content similarity, our approach leverages ontology-based semantic feature engineering and graph-based learning for robust and interpretable predictions. Experimental results show that random forest and graph neural networks significantly outperform traditional models, achieving high accuracy and ranking precision. This work contributes to knowledge management by enabling expert recommendation, trend identification, and semantic integration for strategic academic planning.
Volume: 41
Issue: 3
Page: 1040-1048
Publish at: 2026-03-10
1 2 3
...
1982 1983
Show 6 of 1983

Discover Our Library

Embark on a journey through our expansive collection of articles and let curiosity lead your path to innovation.

Explore Now
Library 3D Ilustration