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

28,428 Article Results

Early detection of food safety risks using BERT and large language models

10.11591/ijeecs.v39.i3.pp1683-1692
Mohammed El Amin Gasbaoui , Soumia Benkrama , Mostefa Bendjima
Sentiment analysis can be a powerful tool in safeguarding public health. This allows authorities to investigate and take action before a foodborne illness outbreak spreads. This paper introduces a novel system that proactively empowers restaurants to identify potential food safety hazards and hygiene regulation violations. The system leverages the power of natural language processing (NLP) to analyze Arabic restaurant reviews left by customers. By fine-tuning a pre-trained BERT mini-Arabic model on three targeted datasets: Sentiment Twitter Corpus, an Algerian dialect dataset, and an Arabic restaurant dataset, the system achieves an impressive accuracy of 91%. Additionally, the system caters to spoken feedback by accepting audio reviews. We utilized Whisper AI for accurate text transcription, followed by classification using a fine-tuned Gemini model from Google on Algerian local comments and others generated using large language models (LLMs) through few-shot learning techniques, reaching an accuracy of 93%. Notably, both models operate independently and concurrently. Leveraging RESTful APIs, the system integrates the solved sub-solutions from each microservice into a fusion layer for a comprehensive restaurant evaluation. This multifaceted approach delivers remarkable results for both modern standard Arabic (MSA) and the Algerian dialect, demonstrating its effectiveness in addressing restaurant food safety concerns.
Volume: 39
Issue: 3
Page: 1683-1692
Publish at: 2025-09-01

Optimizing supervised learning model for thermal comfort and air quality

10.11591/ijeecs.v39.i3.pp1795-1806
Hidayatus Sibyan , Hermawan Hermawan , Ely Nurhidayati
Thermal comfort and indoor air quality are essential factors that directly influence occupants’ health and activity efficiency. Ensuring optimal thermal conditions also supports energy-efficient buildings by preventing energy waste. Machine learning models have been extensively applied to classify thermal comfort and air quality, with supervised learning algorithms such as support vector machine (SVM) and K-nearest neighbor (KNN) showing high accuracy. However, no prior study has compared or combined these two models for simultaneous prediction of thermal comfort and air quality, especially in diverse geographical settings. This study aims to develop and compare SVM and KNN to determine the most accurate model for enhancing thermal comfort and air quality in highland and lowland Islamic boarding schools. Using a quantitative approach, we collected datasets from schools in Wonosobo (highland) and Pontianak (lowland). The results show that KNN outperforms SVM in accuracy, precision, and F1-score. Additionally, a hybrid model integrating both algorithms further improves accuracy, achieving 91%. These findings highlight the effectiveness of machine learning in optimizing environmental conditions in educational settings.
Volume: 39
Issue: 3
Page: 1795-1806
Publish at: 2025-09-01

Security challenges and strategies for CNN-based intrusion detection model for IoT networks

10.11591/ijeecs.v39.i3.pp2012-2022
Wan Fariza Wan Abdul Rahman , Nurul Taqiah Ab Aziz
The rapid proliferation of internet-of-things (IoT) networks has revolutionized various industries but has also exposed them to a myriad of security threats. These networks are particularly vulnerable to sophisticated cyber-attacks due to their distributed nature, resource constraints, and the diverse range of connected devices. To safeguard IoT systems, intrusion detection systems (IDS) have emerged as a critical security measure. Among these, convolutional neural network (CNN)-based models offer promising capabilities in recognizing and mitigating malicious activities within IoT environments. This paper addresses the security challenges specific to IoT networks and explores the critical aspects of identifying malicious packets that threaten their integrity. It also delves into the general challenges associated with implementing IDS in IoT settings, such as the need for real-time detection, resource efficiency, and adaptability to evolving threats. The discussion extends to potential strategies for enhancing CNN-based IDS. The paper concludes by summarizing the key findings and proposing directions for future research to overcome the identified challenges, ultimately contributing to the development of more robust and effective IDS solutions for securing IoT networks.
Volume: 39
Issue: 3
Page: 2012-2022
Publish at: 2025-09-01

Combination of MLF-VO-F and loss functions for VOE from RGB image sequence using deep learning

10.11591/ijeecs.v39.i3.pp1571-1586
Van-Hung Le , Huu-Son Do , Thi-Ha-Phuong Nguyen , Van-Thuan Nguyen , Tat-Hung Do
Visual odometry estimation (VOE) is important in building navigation and pathfinding systems. It helps entities find their way and estimate paths in the environment. Most of the computer vision (CV)-based VOE models are usually evaluated and compared on the KITTI dataset. Multi-layer fusion framework (MLF-VO-F) has had good VOE results from red, green, and blue (RGB) image sequence in Jiang et al. study, using the DeepNet to extract the low-level textures, edges, and deeper high-level semantic features for estimating motion between consecutive frames. This paper proposed a combined model of MLFVO-F as a backbone and loss functions (LFs) (LMSE, LMSE−L2, LCE, and Lcombi) to optimize and supervise the training process of the VOE model. We evaluated and compared the effectiveness of LFs for VOE based on the KITTI and TQU-SLAM datasets with the original MLF-VO-F. From there, choose the appropriate LF combined with the backbone for VOE. The evaluation results on the KITTI dataset show that LCE(RT E is 0.075m, 0.06m on the Seq. #9, Seq. #10, respectively), and Lcombi (trel is 2.21%, 2.67%, 3.59%, 1.01%, and 4.62% on the Seq. #4, Seq. #5, Seq. #6, Seq. #7, Seq. #10, respectively) have the lowest errors and LMSE has the highest errors (AT E is 133.36m on the Seq. #9).
Volume: 39
Issue: 3
Page: 1571-1586
Publish at: 2025-09-01

Microservices caching for container-based IoT system in the edge and cloud

10.11591/ijeecs.v39.i3.pp1652-1660
Rawaa Qasha , Haleema Sulyaman
Microservices enable agile development by dividing internet of things (IoT) programs into autonomous components, ensuring fault tolerance and parallel operation for enhanced productivity. Their adaptability across diverse service types and applications improves IoT system performance. On the other hand, the container is the preferred solution for microservices-based enterprises. To improve the effectiveness of the deployment system presented in our paper 1, we developed a new caching technique to significantly optimize the performance of the deployment system and automate the sharing and re-using of ready-to-run microservices that have been packaged as Docker images. The new caching techniques are seamlessly integrated with our deployment system to optimize the microservices caching of the IoT application by utilizing Docker-based container virtualization and Redis for consistent data sharing. In addition, DevOps and versioning tools such as GOCD and GitHub are integrated into our system to enhance the automatic deployment of the microservices resulting in self-contained, portable, and repeatable IoT microservices. The effectiveness of the proposed techniques is evaluated via various experiments implemented in various working environments where the results show reduced deployment time and the effort required to re-execute the microservices, in addition to the reduction of burden and error that occur when adopting a manual deployment.
Volume: 39
Issue: 3
Page: 1652-1660
Publish at: 2025-09-01

Mechanized network based cyber-attack detection and classification using DNN-generative adversarial model

10.11591/ijeecs.v39.i3.pp1755-1764
Katikam Mahesh , Kunjam Nageswara Rao
These days almost everything is internet. Cyberattacks are the world's most pressing issues. Due to these attacks, Computer systems can be rendered inoperable, disrupted, destroyed or controlled via cyberattacks. Additionally, they can be used to steal, modify, erase, block, or alter data. Most organizations are facing this Issue and lose financially as well as in data security, there are numerous conventional intrusion detection systems (IDS) and firewalls are illustrations for network security tools which are not able to classify and detect different types of attacks in network. With machine learning approach using the Dataset KDD_CUP 99 as input, the synthetic minority oversampling technique (SMOTE) is one of the most often used oversampling methods for addressing imbalance issues. The proposed hybrid deep neural network (DNN), generative adversarial network (GAN), and exhaustive feature selection (EFS) can detect and classify several attack types including R2L, U2R, Probe, denial of service (DoS), and normal attacks types and inform to administrator to ring alarm sound to control and monitor network traffic in dynamically typed networks.
Volume: 39
Issue: 3
Page: 1755-1764
Publish at: 2025-09-01

Design and optimization strategy of HAWT using a local pitch angle adaptation

10.11591/ijeecs.v39.i3.pp1467-1479
Issam Meghlaoui , Toufik Madani Layadi
In this paper, a smart design of a horizontal wind turbine (HAWT) has been developed. The developed design allows improving kinetic energy recuperation with high efficiency. The considered design of the wind turbine is characterized by a specific mechanical structure of blades. Each blade contains separated elements with an adaptable local pitch angle. To develop the smart wind turbine, a new algorithm for controlling the blade elements has been implemented. It allows estimating the distribution of the twist angle of each blade element. The achieved twist angle corresponds to the extracted optimal power provides by the wind turbine. The obtained results show a significant improvement relative to the wind turbine power coefficient under operating conditions. In fact, for some rotating velocity, the rate of this coefficient is increased by 21%. Moreover, notable kinetic energy recuperation is observed. Furthermore, smart orientation of elements proved optimal energy recuperation for a large scale of tip speed ratio and wind speed. In addition, the proposed structure of the wind turbine is more beneficial to minimize the axial thrust. Furthermore, the axial thrust of the wind turbine has been decreased by 21% for some operating velocity and specific conditions. As perspectives for the future works many ideas are suggested.
Volume: 39
Issue: 3
Page: 1467-1479
Publish at: 2025-09-01

IoT-based real-time monitoring of river water quality: a case study of the Selangor River

10.11591/ijeecs.v39.i3.pp1541-1552
Nur Aqilah Ahmad Jafri , Arni Munira Markom , Yusrina Yusof , Norhafizah Burham , Marni Azira Markom
Monitoring river water quality is crucial for preserving freshwater ecosystems, ensuring public health, and supporting resource management. Traditional methods, while accurate, lack the scalability and real-time capabilities needed for proactive intervention. This study introduces an IoT based water quality monitoring system for the Selangor River, integrating sensors for pH, temperature, turbidity, and total dissolved solids (TDS) with a NodeMCU ESP32 microcontroller. To complement the IoT system, a handheld test pen was used to measure salinity and electrical conductivity (EC), offering additional insights into water quality. Field tests at four stations along the river revealed significant spatial variations. Station 1, near the river mouth, showed high salinity, EC, and TDS, indicating saltwater intrusion, with relatively low turbidity. Stations 2 and 3 recorded the highest turbidity levels, suggesting sedimentation and upstream activities, with moderate salinity and EC. Station 4, upstream, demonstrated stable freshwater characteristics, with low salinity, EC, and turbidity levels. The IoT system reliably monitored real-time parameters, and its measurements were validated against those from the handheld test pen. Minor discrepancies in TDS and temperature readings highlighted the importance of calibration.
Volume: 39
Issue: 3
Page: 1541-1552
Publish at: 2025-09-01

A hybrid approach to behavioral spam review detection on e-commerce platforms using apriori and CNN

10.11591/ijeecs.v39.i3.pp1837-1845
Ganesh Wayal , Vijay Bhandari
Spam reviews significantly undermine the credibility of online review systems on e-commerce websites. This paper presents a hybrid methodology that combines the Apriori algorithm and convolutional neural networks (CNN) to efficiently identify and mitigate spam reviews. By examining user behavior, including activity patterns, reviewer reputation, temporal dynamics, and sentiment consistency, we propose a comprehensive model for understanding user interactions and engagement. To extract important information and build precise spam detection models, we use data mining and machine learning approaches. Furthermore, contextual and domain-specific analyses are conducted to improve detection strategies. The study highlights the significance of hybrid techniques in preserving the integrity of e-commerce platforms through successful industry implementations and presents evaluation metrics, problems, and future research objectives.
Volume: 39
Issue: 3
Page: 1837-1845
Publish at: 2025-09-01

CNN-GRU based cyber-attack classification and detection with the CICIDS-2017 dataset using optimization algorithm for honey badger

10.11591/ijeecs.v39.i3.pp1765-1775
Katikam Mahesh , Kunjam Nageswara Rao
The sheer volume of data exchanged has grown through information and communications technology (ICT) swiftly growing importance since the attackers benefit from illegal access to network data and introduce possible dangers for data theft or alteration. It is considered a significant barrier to monitor the network traffic for cyber-attack detection and classification with alarm ring to inform to network administrator. With KDD-CUP99, conventional machine learning methods like deep neural network (DNN), a kind of artificial neural network (ANN), cannot detect and classify novel attacks types and lacks clarity regarding accuracy. The CICIDS 2017 dataset, which is improved in this study, serves as training data for the model and useful framework that combines a hybrid convolutional neural network (CNN) with the gated recurrent unit (GRU) technique. The primary aim of this effort is to classify different security attacks and classify cyberthreats with honey badger optimization algorithm (HBOA). To strengthen the performance criteria for various assault types, such as F1-score, recall, precision, and others, the HBOA is utilized to modify the model parameters high-level features ought to be extracted from the network data using the hybrid model assessed and verified by simulation studies. The detection and classification output from the CNN-GRU model, which detects different security threats with greater accuracy of 94%.
Volume: 39
Issue: 3
Page: 1765-1775
Publish at: 2025-09-01

Modelling and simulation of maximum power point tracking on partial shaded PV based-on a physical phenomenon-inspired metaheuristic algorithm

10.11591/ijeecs.v39.i3.pp1923-1937
Prisma Megantoro , Joy Sefine Dona Saya , Muhammad Akbar Syahbani , Marwan Fadhilah , Pandi Vigneshwaran
Maximum power point tracking (MPPT) is a technique to optimize the photovoltaic (PV) current generation, so it can improve the efficiency of solar energy harvesting. MPPT works by searching the voltage which generates the maximum power, called the maximum power point (MPP). MPP value changes by the fluctuance of ambient temperature and solar insolation level depicted by the I-V curve. Searching the MPP will be more complex if the partial shading is happened. The effect of partial shading will rise to more than one local MPPs. In this research, an optimization algorithm is modeled and simulated the MPPT technique in partial shading. The optimization uses the new metaheuristic algorithm which inspired from a physical phenomenon, called Archimedes optimization algorithm (AOA). The AOA uses mathematical modeling which has convergence capabilities, balanced exploration, and exploitation and is suitable for solving complex optimization technique, like MPPT. The research used varies partial insolation percentage. The implementation of MPPT-AOA compared to other metaheuristic algorithms to analysis its performance in the aspect of PV system parameters and tracking process parameters. The simulation result shows that the AOA can enrich the MPPT technique and improve the solar energy harvesting which is superior to other algorithms.
Volume: 39
Issue: 3
Page: 1923-1937
Publish at: 2025-09-01

Optimizing energy efficiency and improved security in wireless sensor networks using energy-centric MJSO and MACO for clustering and routing

10.11591/ijeecs.v39.i3.pp1964-1975
Srinivas Kalaskar , Channappa Bhyri
Wireless sensor networks (WSNs) play a pivotal role in various applications, but their energy-constrained nature poses significant challenges to their sustainable operation. In this paper, we propose a novel approach to enhance energy efficiency in WSNs by leveraging energy-centric multi-objective jaya search optimization (MJSO) and multi-objective ant colony optimization (MACO) for clustering and routing. Our method aims to address the energy consumption issues by optimizing clustering and routing strategies simultaneously. The energy-centric MJSO algorithm is employed to intelligently organize sensor nodes into clusters, considering energy consumption, network coverage, and connectivity. The multi-objective MACO algorithm optimizes routing paths by balancing energy consumption and network lifetime objectives. Through integration and simulations, the approach enhances energy efficiency in WSNs for various applications like environmental monitoring and smart cities, advancing energy-efficient clustering and routing. By integrating energy-centric MJSO and MACO into clustering and routing protocols, WSNs can achieve significant improvements in energy efficiency and security while maintaining reliable communication and data delivery.
Volume: 39
Issue: 3
Page: 1964-1975
Publish at: 2025-09-01

Unveiling critical factors of test automation adoption in software testing

10.11591/ijeecs.v39.i3.pp1826-1836
Miftahul Kahfi Al Fath , Tanty Oktavia
This paper aims to observe the adoption of test automation in Indonesia and examine the determining factors that influence the use of this technology in organizations. The study focuses on five critical factors: technology acceptance model, task-technology fit, managerial support (MS), individual performance, and organizational performance. A survey of 109 QA community members was conducted to collect data, and partial least squares structural equation modeling was used for data processing. Based on the study, Selenium is the top test automation framework used for organizations in Indonesia, followed by Appium and Postman. The result showed that out of twelve (12) examined relationships, nine (9) of them were accepted. This data indicates the strong influence of task technology fit (TTF), computer self-efficacy (CSE), perceived ease of use, perceived usefulness, and MS towards behavioral intention and actual use of test automation. Additionally, the actual use of test automation was found to have a positive impact on individual and organizational performance. The study contributes valuable insights for decision-makers by identifying critical factors influencing automation adoption and offers a replicable methodology for evaluating similar technologies.
Volume: 39
Issue: 3
Page: 1826-1836
Publish at: 2025-09-01

Texture-based two-stage shot boundary detection in videos

10.11591/ijeecs.v39.i3.pp1955-1963
S. Anitha , J. Kavitha , G. Prince Devaraj , Shachi Mall , S. Suma Christal Mary , Ezhil R
In recent years, shot boundary detection (SBD) has become an essential component of video processing, enabling applications such as video indexing, summarization, and content retrieval. However, the task remains challenging due to frequent false positive detections caused by illumination variations, motion changes, and diverse editing effects. To address these challenges, this paper presents a novel two-stage SBD framework that leverages local quad pattern (LQP) histogram features for precise transition detection. In the first stage, histogram feature vectors are derived by counting the occurrences of LQP codes (−1, +1, 1, 0), and abrupt transitions are identified using the Euclidean distance between consecutive frames. In the second stage, mean values of each histogram bin are computed for consecutive frames, and a similar distance-based approach is applied to refine detection accuracy. A transition frame is confirmed as a shot boundary only if both stages detect it, thereby reducing false positives. The proposed method is evaluated on the TRECVid 2001 and 2007 benchmark datasets, and experimental results demonstrate its superior performance compared to existing algorithms.
Volume: 39
Issue: 3
Page: 1955-1963
Publish at: 2025-09-01

Prediction of Parkinson's disease using feature selection and ensemble learning techniques

10.11591/ijeecs.v39.i3.pp1736-1744
Sharan T. D. , Sujata Joshi
Parkinson's disease (PD) is a progressive neurodegenerative disorder that significantly impacts quality of life and healthcare systems. Early detection is crucial for timely interventions that can mitigate disease progression and improve patient outcomes. This study leverages advanced machine learning (ML) techniques to detect PD using speech features as non-invasive biomarkers. A dataset containing 754 features derived from sustained vowel phonations of 252 individuals (188 PD patients, 64 healthy controls) was analyzed. The dataset, originally collected by Istanbul University and publicly hosted via the UCI ML repository, was accessed through Kaggle for preprocessing and analysis. To identify the most predictive features, we employed recursive feature elimination (RFE), random forest importance, lasso regression, and the boruta algorithm—ensuring robust feature selection while reducing dimensionality. The XGBoost model, optimised using synthetic minority oversampling technique (SMOTE) for class balancing, achieved an accuracy of 96.69%, a recall of 96%, and an F1-score of 98%. Model robustness was validated through 5-fold cross-validation, yielding an average accuracy of 89.54%. These findings establish a scalable, costeffective, and non-invasive framework for early PD detection, demonstrating the potential of speech analysis and ML in neurodegenerative disease management.
Volume: 39
Issue: 3
Page: 1736-1744
Publish at: 2025-09-01
1 2 3
...
1895 1896
Show 32 of 1896

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