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

30,938 Article Results

Margin-reciprocal loss: enhancing robust network anomaly detection on imbalanced traffic data

10.11591/ijeecs.v42.i2.pp498-508
Rachid Tahri , Abdellah Ouammou , Abdellatif Lasbahani
Accurate detection of network intrusions remains challenging under severe class imbalance, where rare attacks such as remote-to-local (R2L) and user-to-root (U2R) are poorly represented. Although many learning-based intrusion detection systems achieve high overall accuracy, conventional loss functions often bias training toward majority classes, leading to weak minority-class performance. This paper introduces a smooth margin-reciprocal loss (MRL), inspired by distance-weighted discrimination (DWD), which emphasizes samples with small or negative margins while rapidly attenuating penalties for well-classified instances. Unlike probability-based focal loss, MRL operates directly on the signed margin and enables stable optimization with first-order methods. Experiments conducted on the NSL-KDD benchmark using linear and shallow multilayer perceptron models show that MRL consistently improves macro-F1 and per-class precision–recall AUC compared with hinge, logistic, and focal losses, with notable gains on minority attack classes.
Volume: 42
Issue: 2
Page: 498-508
Publish at: 2026-05-10

A hybrid large language model-graph neural network framework for Arabic sentiment analysis

10.11591/ijeecs.v42.i2.pp391-402
Hani Mohammadn Iwidat
Arabic sentiment analysis (SA) faces significant challenges due to the language’s morphological richness and dialectal diversity. This study introduces a novel hybrid large language model-graph neural network (LLM-GNN) framework designed to address these challenges. The proposed model integrates the contextual understanding of AraBERT v2 with the structural learning capability of a graph convolutional network (GCN). It constructs a graph of sentences using cosine similarity, allowing the GCN to capture crucial inter-sentence semantic dependencies often missed by sequential models. Findings: The model is evaluated on a publicly available Arabic 100k Reviews dataset consisting of authentic user-generated Arabic reviews balanced across positive, negative, and mixed sentiment classes. The results demonstrate that the proposed LLM-GNN model performed better as compared to the baseline models, including fine-tuned AraBERT, AraBERT-BiLSTM, AraBERT-MLP, and multilingual BERT. The hybrid model achieves an overall accuracy of 66.8% and a F1-score of 66.55%, with an improvement of 7.6% and 4.4%, respectively. The model demonstrated stable convergence from the first training epoch. Research limitations/implications: The graph construction is performed at the mini batch level, which restricts the modeling of global semantic relationships across the entire corpus. The results show that the hybrid model identifies subtle sentiment cues that sequential models frequently miss by fusing relational graph reasoning with contextual embeddings. By effectively identifying subtle sentiment cues, the hybrid model can significantly enhance the accuracy of real-world applications such as social media monitoring and customer review analysis for Arabic content.
Volume: 42
Issue: 2
Page: 391-402
Publish at: 2026-05-10

Greenhouse irrigation system based on AIoT

10.11591/ijeecs.v42.i2.pp531-541
Tariq Benahmed , Seddiki Noureddine , Benahmed Khelifa
Agriculture is one of the biggest consumers of fresh water. Different types of irrigation systems are available and used in agricultural greenhouses. These irrigation systems are traditional, which do not allow for water savings and have very high maintenance costs. Smart irrigation in agricultural greenhouses is an innovative approach that contributes to a more efficient use of water in agriculture by optimizing available resources and promoting a more sustainable and profitable production. Through the use of smart technologies such as the artificial intelligence internet of things (AIoT), smart irrigation allows adjusting the amounts of water provided to the actual needs of plants, based on factors from sensors such as temperature, air and soil humidity, crop growth stages, plant conditions, and soil types. This helps to avoid water waste and reduce the risks of water stress for plants, while improving the quality and yield of crops. This article presents a smart irrigation solution in greenhouses whose results have been validated by an experimental prototype.
Volume: 42
Issue: 2
Page: 531-541
Publish at: 2026-05-10

Development of an adaptive student behavior model for e tutoring systems

10.11591/ijeecs.v42.i2.pp561-571
Rania Ali Elkhidir Ali , Hussein Ali Ahmed Ghanim , Maha Osman , Nazar Faried Yousif Mohamed
Static e-tutoring systems typically utilize rigid educational sequences that do not adapt to learners' changing knowledge states, engagement levels, and cognitive requirements. This constraint frequently leads to ineffective learning and heightened cognitive strain. This paper presents an integrated adaptive student behavior model (ASBM) that tackles this challenge by functioning at the granularity of interaction steps. It integrates bayesian knowledge tracing (BKT) for probabilistic skill mastery assessment, an LSTM-based deep neural network for behavioral feature extraction, and a deep Q-network for adaptive pedagogical decision-making. The proposed methodology underwent evaluation via a randomized controlled experiment with 120 undergraduate students over a three-week educational duration. Participants were allocated to either an adaptive E-Tutoring system utilizing an integrated ASBM or to a static, non-adaptive system. The quantitative results indicate that the adaptive system attained a superior normalized learning gain (0.72 compared to 0.57, p < 0.01), reduced time to mastery (45 minutes vs 65 minutes), enhanced delayed retention (+18%), elevated completion rates (92% versus 78%), and diminished subjective cognitive burden. The results demonstrate that fine-grained adaptivity, facilitated by a hybrid bayesian knowledge tracing, deep neural network, and reinforcement learning (RL) architecture, markedly improves learning efficiency and learner experience in controlled experimental settings. The research provides empirical evidence that supports the amalgamation of cognitive and behavioral modeling with reinforcement learning for advanced e-tutoring systems.
Volume: 42
Issue: 2
Page: 561-571
Publish at: 2026-05-10

Meta-stacking models for electricity load forecasting in West Java

10.11591/ijeecs.v42.i2.pp442-453
Denanda Aufadlan Tsaqif , Bagus Sartono , Hari Wijayanto
Indonesia’s electricity demand continues to increase due to population growth, urbanization, and industrial expansion, therefore making accurate load forecasting is essential to maintain supply-demand balance. However, electrical load demand in West Java has a complex pattern (seasonality, nonlinear behavior, weather variability, and holiday effects), which motivates the use of a meta-stacking approach to effectively capture such complexity. Previous research shows that meta-stacking outperforms individual models, but it fails to capture sudden changes and its performance consistency remains unclear. Therefore, this study proposes a meta-stacking framework for daily electricity load forecasting in West Java (2006-2023) that includes weather and holiday variables by combining CNN-BiLSTM, CNN-BiGRU, and Windowed-XGBoost forecasts through linear regression and evaluates its performance across five data-splitting scenarios and nine forecast horizons, which represents the main novelty in this research. Meta stacking shows strong generalization across scenarios and strong long-term forecasting performance across horizons, while consistently providing a balanced trade-off between MAPE and trend accuracy, where the model trained on the longest historical dataset achieves the best performance with 1.89% MAPE and 86% trend accuracy. The proposed approach successfully captures seasonal and holiday-related load patterns, indicating its potential to support PLN in improving demand planning and operational decision making.
Volume: 42
Issue: 2
Page: 442-453
Publish at: 2026-05-10

Natural language processing for report consolidation and matching based on latent semantic analysis and cosine similarity

10.11591/ijeecs.v42.i2.pp609-618
Jeleen M. Mangubat , Ryndel Ventura Amorado , Lovely Rose T. Hernandez , Jennifer L. Marasigan
Consolidation of reports and matching of documents pose several challenges especially when dealing with large amounts of textual data. Thus, organizations are in need of intelligent systems that are capable of automating these processes, ensuring faster, more accurate analysis and retrieval of relevant information. This study applies Latent Semantic Indexing (LSI) and Cosine Similarity to automate the matching of gender related issues, activities, and programs submitted by university offices. An intelligent web-based system was developed using Python and Django to implement these algorithms for report consolidation. Performance evaluation using accuracy, precision, recall, and F1-score demonstrated that the model correctly classified 90% of entries. A threshold sweep experiment further revealed that a similarity value of 0.51 provides the optimal decision boundary for identifying semantically similar instances. The findings confirm that LSI remains effective for low-resource institutional text analysis, enabling more efficient and accurate report consolidation.
Volume: 42
Issue: 2
Page: 609-618
Publish at: 2026-05-10

GRAND-stream: A galois-ring-based lightweight stream cipher for battery-limited internet of things (IoT) devices

10.11591/ijeecs.v42.i2.pp542-551
Nahom Gebeyehu Zinabu , Yihenew Wondie Marye , Kula Kekeba Tune , Samuel Asferaw Demilew
GRAND-Stream is a new lightweight stream encryption framework based on arithmetic over Galois rings that targets battery-constrained internet of things (IoT) devices. Unlike traditional LFSR/NLFSR-based designs, GRAND-stream uses ring-squaring-induced nonlinearity and inter component polynomial coupling to improve algebraic complexity while keeping compact implementation qualities. We give an explicit parameterized construction for Z2 n[x]/(f(x)), specify its state updating and output functions, and investigate algebraic degree growth and diffusion behavior. The security arguments are heuristic, based on explicitly stated assumptions about the difficulty of solving quadratic systems over Galois rings. Energy per bit, cycle count, and gate complexity are estimated using an analytical performance model. Although initial findings show potential compactness, more research is needed for thorough cryptanalysis and empirical validation on embedded devices. Therefore, GRAND-stream should be considered a structured algebraic design concept that needs more assessment.
Volume: 42
Issue: 2
Page: 542-551
Publish at: 2026-05-10

A lightweight architecture for IoT based on blockchain, designed for constrained IoT devices

10.11591/ijeecs.v42.i2.pp596-608
Yassin Elgountery , Mohamed Aghroud , Meryem Lasaad , Mohamed Oualla
Blockchain is a technology that is evolving day by day, characterized by features such as security, decentralization, immutability, traceability, and privacy protection. These features make it a promising solution for internet of things (IoT) systems. However, the inherent constraints of IoT devices in terms of storage, computation, energy capacity, and other aspects present significant challenges to integrating blockchain into these systems. This emphasizes the necessity of developing a lightweight solution that considers these specific constraints. This conceptual article proposes a lightweight architecture based on delegated nodes, centered on blockchain technology, and an optimized practical byzantine fault tolerance (PBFT) consensus algorithm, to ensure scalability and reliability for IoT. Moreover, to reduce the storage overhead in the blockchain, an off-chain cloud-based storage solution is proposed in this article. The proposed architecture is designed to prevent direct IoT device-blockchain interactions. All system operations are defined in a single smart contract, which helps reduce the complexity and overhead of the system.
Volume: 42
Issue: 2
Page: 596-608
Publish at: 2026-05-10

Modeling and control of a solar-powered cable-driven robot under power constraints

10.11591/ijeecs.v42.i2.pp318-336
Kendouli Fairouz , Hemama Aboud , Khoudir Abed
This paper investigates the control performance of a solar-powered cable driven robot using MATLAB simulations, comparing a conventional proportional–integral–derivative (PID) controller with a hybrid fuzzy PD controller. The study adopts a simplified kinematic model to focus on control behavior under cable-induced nonlinearities and time-varying power availability due to solar energy, while neglecting full dynamic and cable tension effects. Both controllers were systematically tuned to ensure a fair comparison. Performance was evaluated in terms of speed and position tracking under fluctuating solar conditions. Simulation results show that the hybrid fuzzy PD controller provides superior performance, with lower RMS tracking errors, reduced overshoot, and faster settling times compared to the PID controller. These findings highlight the potential of energy-aware intelligent control strategies for improving the reliability and accuracy of solar-powered cable-driven robots operating under variable renewable energy conditions.
Volume: 42
Issue: 2
Page: 318-336
Publish at: 2026-05-10

Enhancing NICD and NIMH batteries charging efficiency: a MSCCC strategy using artificial intelligence control

10.11591/ijeecs.v42.i2.pp349-368
Somendra Banerjee , Awdhesh Kumar , Vinod Kumar Giri
In the above essay, a smart multi-stage constant current charging (MSCCC) strategy has been proposed with an adaptive neuro-fuzzy inference system (ANFIS) to improve the charging efficiency of nickel metal hydride (NiMH) and nickel cadmium (NiCd) type batteries. The suggested charger uses a boost converter that is power-factor-corrected and variable current regulation according to real-time feedback of voltage and state of charge. MATLAB/Simulink is used to test the system with a 24 V23.5 Ah NiCd pack and 25.2 V49.4 Ah NiMH pack. Comparative simulations on conventional PI, fuzzy, and neural controllers show that ANFIS-MSCCC approach enhances state-of-charge (SoC) retention by about 5-8 percent, voltage overshoot by almost 20 percent and transitions between currents are smoother which results into lower electrical stress. Besides, the suggested approach has a shorter settling time, high charging stability, and safe thermal characteristics. These findings prove that the ANFIS-aided MSCCC provides a powerful and reconfigurable charging system to NiCd and NiMH batteries, which is applicable within the complex battery management systems that are already in use.
Volume: 42
Issue: 2
Page: 349-368
Publish at: 2026-05-10

Assessing cybersecurity awareness and security practices among university students in Jordan

10.11591/ijeecs.v42.i2.pp619-630
Khader Musbah Ismail Titi
This paper sets out to evaluate the degree to which Jordanian university students genuinely comprehend and apply cybersecurity principles in their everyday digital lives. The rationale for undertaking this investigation is compelling: as cyber threats continue to intensify and diversify, remarkably little granular evidence exists to identify which sub-populations within the Jordanian student body are most vulnerable due to knowledge gaps. A structured survey reaching 150 students recruited from a range of academic departments served as the empirical foundation, with all quantitative analyses conducted using SPSS and MS-Excel. The analysis examined three demographic dimensions: gender, academic discipline, and geographic origin (urban versus rural). Students enrolled in computing and information technology programmes consistently demonstrated superior preparedness relative to their peers, and urban-based students exhibited more robust awareness profiles than those from rural areas. Building on these findings, the paper makes a case for systematic awareness programmes and for embedding cybersecurity content into curricula across all disciplines rather than restricting it to technical fields.
Volume: 42
Issue: 2
Page: 619-630
Publish at: 2026-05-10

Fairness dynamics in graph neural networks: a comparative study of graph-structured neural models with and without gradient-based training

10.11591/ijeecs.v42.i2.pp403-413
Ananda Chatterjee , K A Venkatesh
Graph neural networks (GNNs) are gaining more and more popularity in high stakes domain due to their ability to learn both from features and relationships. Nevertheless, there are concerns regarding how this accuracy centric optimization used by these models will impact fairness when deployed in socially sensitive areas. This work explores the interplay between predictive accuracy and fairness in GNNs when applied in judicial risk assessment system. A comparative study was performed among three canonical architectures such as graph convolutional networks (GCN), graph sample and aggregate (GraphSAGE) and graph attention networks (GAT) under trained and untrained settings on judicial risk assessment dataset. Fairness was evaluated through metrices like demographic parity (DP), equalized opportunity (Eopp), and equalized odds (Eodds) along with predictive performance metrices. Sensitivity analysis was conducted to investigate the effect of graph construction choices and neighborhood sizes in determing fairness and predictive accuracy. Experimental evidences proved that backpropagation improved predictive performance but in tandem fairness degradation happened. Untrained models exhibited lower fairness gap but that is superficial as weak predictive outcome of those models made group differences suppressed. Among the three trained models GAT was able to strike a good balance between accuracy and fairness while increase in neighborhood size caused little bit improvement in fairness via graph smoothing. The novelty of this work lies with its empericial characterization of GNNs under realistic settings. This study emphasizes the fact that how learning methodology, architectural designs, graph formation influence fairness outcomes. This work enlightens how graph-based models can be applied to decision making scenario and encourages embedding of fairness aware training strategies to it.
Volume: 42
Issue: 2
Page: 403-413
Publish at: 2026-05-10

Parametric analysis and mitigation of Ferroresonance in ungrounded power transformers: Algeria power plant approach

10.11591/ijeecs.v42.i2.pp307-317
Mohammed Boukaf , Saliha Boutora , Hamid Bentarzi
Ferroresonance in ungrounded power transformers is extremely dangerous for electric systems, causing serious overvoltage, equipment damage, and system instability. Until now, there have been fewer quantitative, parametric analytical studies and verified mitigation methods for real power systems. Although there has been a qualitative study of Ferroresonance, few have provided rigorous parametric assessments and validated mitigation actions for modern power plants at utility scale. This work fills these gaps through in-depth parametric analysis of the so-measured and so-calculated values and a validation of zigzag transformer grounding for the Ras Djinet 1131 MW power plant, serving as a case study in Algeria. This paper establishes a nonlinear mathematical model that reflects transformer saturation characteristics, grading capacitance (Cg), shunt capacitance (Csh), and magnetization resistance (Rm), and implements it in MATLAB/Simulink. Fundamental and chaotic Ferroresonance modes are identified using parametric sweeps spectral analysis (FFT). Act two: Modeling and validating zigzag grounding transformer mitigation. The study finds Cg (0.5–7 µF) increases strengthen Ferroresonance overvoltage, while Csh (0.3_7µF) and Rm (1500–65860Ω) reduce it, offering a damping effect, reducing peak voltages. Spectral analysis reveals the fundamental mode (dominated by 50 Hz), the Quasi-Periodic Mode, and the chaotic mode (multiple frequency components). With a zero-sequence current path, a zigzag grounding implementation completely removes Ferroresonance in all tested cases.
Volume: 42
Issue: 2
Page: 307-317
Publish at: 2026-05-10

Enhancing IPv6 enabled IoT system security using addressless architecture

10.11591/ijeecs.v42.i2.pp469-484
Ashmita Tiwari , Chitran Pokhrel , Babu R. Dawadi
The growth and sensitivity of internet of things (IoT) deployments demand robust and efficient security mechanisms, especially at the addressing layer. Traditional IPv6 addressing is susceptible to scanning, spoofing, and tracking, especially in IPv6 over low-power wireless personal area networks (6LoWPAN) networks. This paper proposes a dynamic elliptic curve cryptography (ECC)-based IPv6 address generation mechanism for 6LoWPAN IoT networks. Encrypting Interface IDs (IIDs) while keeping the network prefix the same to improve security against scanning, inference, and correlation attacks. High entropy of 0.9836 and cryptanalysis confirm higher randomness and high resistance to wide vectors of attacks. Having computed an average delay of encryption as 2.5728 ms, the process ensures low latency and insignificant overhead. It is more secure and efficient than existing techniques and hence is ideal for real-time resource- constrained IoT applications.
Volume: 42
Issue: 2
Page: 469-484
Publish at: 2026-05-10

Cost-effective sentiment analysis with chain-of-thought: a cross-lingual evaluation

10.11591/ijeecs.v42.i2.pp454-468
Shen Haijie , Madhavi Devaraj
Sentiment analysis is a core task in natural language processing with broad ap plications in social media monitoring, customer feedback mining, and market research. Although pre-trained language models (e.g., BERT) achieve strong performance, they typically rely on task-specific fine-tuning and substantial la beled data. Recent large language models (LLMs) enable a different paradigm via in-context learning. This paper presents a systematic empirical study investi gating chain-of-thought sentiment (CoT-Sent), a prompting framework that uses structured CoT reasoning to improve classification accuracy. We evaluate CoT Sent on four benchmark datasets in English and Chinese, comparing multiple representative LLMs (GPT-4, Claude-3, Gemini, Qwen-2.5) under zero-shot set tings. Across datasets, CoT-Sent improves average accuracy by 2.5% over zero shot baselines. Crucially, unlike prior work which provides a broad performance overview without analyzing deployment costs or multi-language generalization, we focus on the cost-latency-accuracy trade-offs, and demonstrate CoT-Sent’s superior cross-lingual transfer (English-to-Chinese) with detailed cost analysis. We provide a comprehensive three-dimensional analysis of accuracy, cost, and latency, offering actionable deployment strategies for resource-constrained environments.
Volume: 42
Issue: 2
Page: 454-468
Publish at: 2026-05-10
1 2 3
...
2062 2063
Show 63 of 2063

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