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

31,456 Article Results

Target image validation modeling using deep neural network algorithm

10.11591/ijece.v15i2.pp2042-2054
Naemah Mubarakah , Poltak Sihombing , Syahril Efendi , Fahmi Fahmi
Research on image validation models is an interesting topic. The application of deep learning (DL) for object detection has been demonstrated to effectively and efficiently address the challenges in this field. Deep neural networks (DNN) are deep learning algorithms capable of handling large datasets and effectively solving complex problems due to their robust learning capacity. Despite their ability to address complex problems, DNN encounter challenges related to the necessity for intricate architectures and a large number of hidden layers. The objective of this research is to identify the most effective model for achieving optimal performance in image validation. This study investigates target image validation using DNN algorithms, examining architectures with 3, 4, 5, and 6 hidden layers. This study also evaluates the performance of image validation across various activation functions, batch sizes, and numbers of neurons. The results of the study show that the best performance for image validation is achieved using the Leaky-ReLU and Sigmoid activation functions, with a batch size of 64, and an architecture consisting of 3 hidden layers with neuron sizes of 256, 128, and 64. This model is capable of providing real-time target image validation with an accuracy of up to 94.31%.
Volume: 15
Issue: 2
Page: 2042-2054
Publish at: 2025-04-01

Improving water quality parameter prediction with multi-level linear regression model and hybrid feature selection

10.11591/ijece.v15i2.pp2381-2391
Aleefia Khurshid , Samruddhi Korke , Yudhir Kothari , Shruti Alone , Khushali Bais
Predicting and modeling the quality of water is essential to guarantee that the water is safe to drink. The chlorine content in water needs to be monitored in real-time to provide a consistent supply of drinkable water. Additionally, potassium and chlorine have a major impact on how appealing the water is, as they are important components that influence taste and odor. Therefore, to evaluate the levels of chlorine and potassium, this work presents a multivariable linear regression approach backed by a hybrid feature extraction method. To bridge the gap between the filter and wrapper approaches, a hybrid approach is used to remove unnecessary information and reduce processing time and complexity. Here the quantitative parameters, in conjunction with categorical parameters, are instrumental in enabling accurate prediction of two water quality parameters. The two developed multi-level regression (MLR) models for the prediction of potassium and chloride are useful when factors affecting water parameters fluctuate at the site level as well as over larger spatial or temporal scales giving consumers a visual representation of how each parameter influences prediction. The converged model outperforms in comparison with other machine learning algorithms with an MAE of 7.42e-15 for potassium and 3.72e-14 for chloride.
Volume: 15
Issue: 2
Page: 2381-2391
Publish at: 2025-04-01

Intrusion detection based on generative adversarial network with random forest for cloud networks

10.11591/ijece.v15i2.pp2491-2498
Gnanam Jeba Rosline , Pushpa Rani
The development of cloud computing enables individuals and organizations to access a wide range of online programs and services. Because of its nature, numerous users can access and distribute cloud infrastructure. In cloud computing several security threats change the data and operations. A network's ability to detect malicious activity and possible threats is greatly aided by intrusion detection. To solve these issues, intrusion detection based on generative adversarial network with random forest (GAN-RF) for cloud networks is introduced. The function of the generative adversarial networks (GANs) based network abnormality recognition system is evaluated. It uses the CICIDS2018 dataset to detect intrusion. GAN is utilized to improve network anomaly detection in conjunction with an ensemble random forest (RF) classifier. The GAN-RF model achieved 95.01% of accuracy for intrusion detection and obtain better recall and F1-score. Extensive assessments and valuations illustrate the efficiency of the GAN-RF approach in accurately identifying network issues.
Volume: 15
Issue: 2
Page: 2491-2498
Publish at: 2025-04-01

Advancing network security: a comparative research of machine learning techniques for intrusion detection

10.11591/ijece.v15i2.pp2271-2281
Shynggys Rysbekov , Abylay Aitbanov , Zukhra Abdiakhmetova , Amandyk Kartbayev
In the current digital era, the advancement of network-based technologies has brought a surge in security vulnerabilities, necessitating complex and dynamic defense mechanisms. This paper explores the integration of machine learning techniques within intrusion detection systems (IDS) to tackle the intricacies of modern network threats. A detailed comparative analysis of various algorithms, including k-nearest neighbors (KNN), logistic regression, and perceptron neural networks, is conducted to evaluate their efficiency in detecting and classifying different types of network intrusions such as denial of service (DoS), probe, user to root (U2R), and remote to local (R2L). Utilizing the national software laboratory knowledge discovery and data mining (NSL-KDD) dataset, a standard in the field, the study examines the algorithms’ ability to identify complex patterns and anomalies indicative of security breaches. Principal component analysis is utilized to streamline the dataset into 20 principal components for data processing efficiency. Results indicate that the neural network model is particularly effective, demonstrating exceptional performance metrics across accuracy, precision, and recall in both training and testing phases, affirming its reliability and utility in IDS. The potential for hybrid models combining different machine learning (ML) strategies is also discussed, highlighting a path towards more robust and adaptable IDS solutions.
Volume: 15
Issue: 2
Page: 2271-2281
Publish at: 2025-04-01

Squeeze-excitation half U-Net and synthetic minority oversampling technique oversampling for papilledema image classification

10.11591/ijai.v14.i2.pp1410-1419
Wiharto Wiharto , Angga Exca Pradipta Syaifuddin
The emergence of various convolutional neural networks (CNN) architectures indicates progress in the computer vision field. However, most of the architectures have large parameters, which tends to increase the computational cost of the training process. Additionaly, imbalanced data sources are often encountered, causing the model to overfit. The aim of this study is to evaluate a new method to classify retinal fundus images from imbalanced data into the corresponding classes by using fewer parameters than the previous method. To achieve this, squeeze-excitation half U-Net (SEHUNET) architecture, a modification of half U-Net with squeeze-excite process to provide attention mechanism on each feature maps channel of the model, in combination with synthetic minority oversampling technique (SMOTE) is proposed. The test accuracy of SEHUNET is 98.52% with area under the curve of receiver operation characteristic (AUROC) of 0.999. This result outperforms the previous study that used CNN with Bayesian optimization, achieving accuracy of 95.89% and AUROC of 0.992. SEHUNET is also able to compete with the transfer learning methods used in previous research such as InceptionV3 with 96.35% accuracy, visual geometry group (VGG) with 96.8%, and ResNet with 98.63%. This performance can be achieved by SEHUNET with only 0.268 million parameters compared to the architecture parameters used in previous research ranging from 11 million to 33 million.
Volume: 14
Issue: 2
Page: 1410-1419
Publish at: 2025-04-01

Explainable zero-shot learning and transfer learning for real time Indian healthcare

10.11591/ijict.v14i1.pp91-101
Swati Saigaonkar , Vaibhav Narawade
Clinical note research is globally recognized, but work on real-time data, particularly from India, is still lagging. This study initiated by training models on medical information mart for intensive care (MIMIC) clinical notes, focusing on conditions like chronic kidney disease (CKD), myocardial infarction (MI), and asthma using the structured medical domain bidirectional encoder representations from transformers (SMDBERT) model. Subsequently, these models were applied to an Indian dataset obtained from two hospitals. The key difference between publicly available datasets and real-time data lies in the prevalence of certain diseases. For example, in a real-time setting, tuberculosis may exist, but the MIMIC dataset lacks corresponding clinical notes. Thus, an innovative approach was developed by combining a fine-tuned SMDBERT model with a customized zero-shot learning method to effectively analyze tuberculosis-related clinical notes. Another research gap is the lack of explainability because deep learning (DL) models are inherently black-box. To further strengthen the reliability of the models, local interpretable model-agnostic explanations (LIME) and shapley additive explanations (SHAP) explanations were projected along with narrative explanations which generated explanations in a natural language format. Thus, the research provides a significant contribution with ensemble technique of zero-shot learning and SMDBERT model with an accuracy of 0.92 as against the specialized models like scientific BERT (SCIBERT), biomedical BERT (BIOBERT) and clinical BioBERT.
Volume: 14
Issue: 1
Page: 91-101
Publish at: 2025-04-01

Data analysis and visualization on titanic and student’s performance datasets-an exploratory study

10.11591/ijict.v14i1.pp68-76
Seong-Cheol Kim , Surender Reddy Salkuti , Alka Manvayalar Suresh , Madhu Sree Sankaran
Exploratory data analysis (EDA) is all about exploring the data in order to identify any underlying pattern before you try to use it to make a predictive model. It also plays a major role in the data discovery process as it is used to analyze data and to recapitulate their different characteristics, which is displayed efficiently with the help of data visualization methods. This paper aims to identify errors in the dataset, to understand the existing hidden structure and to identify new ones, to detect points in a dataset that deviate to a greater extent from the collected data (outliers), and also to find any relationship or intersection between the variables and constants. Two datasets are used namely ‘Titanic’ and ‘student’s performance’ to perform data analysis and ‘data visualization’ to depict ‘exploratory data analysis’ which acts as an important set of tools for recognizing a qualitative understanding. The datasets were explored and hence it assisted with identifying patterns, outliers, corrupt data, and discovering the relationship between the fields in the dataset.
Volume: 14
Issue: 1
Page: 68-76
Publish at: 2025-04-01

Improved convolutional neural network-based bearing fault diagnosis using multi-phase motor current signals

10.11591/ijece.v15i2.pp1656-1669
Hai Dang Huu , Ngoc-My Bui , Van-Phuc Hoang , Thang Bui Quy , Yen Hoang Thi
Diagnosing bearing faults of the induction motor is crucial for the maintenance of rotating electrical machines. Numerous methods have been developed and published for monitoring and classifying these faults using sensor data such as vibration, audio, and current signals. Ideally, the current phases are balanced; however, faults disrupt this symmetry, causing each phase to reveal unique diagnostic details. Consequently, studies that rely on a single phase of the current signal may not capture all fault-related characteristics. Research on motor bearing fault diagnosis using two current phases typically extracts features from each phase separately, applying machine learning to classify the faults. Currently, no approach has been proposed to extract features from both phases simultaneously. Furthermore, the proposed solutions have only been published with noise-free data. To address these challenges, this paper introduces an enhanced solution that improves the accuracy of motor bearing fault classification based on an improved convolutional neural network that processes current signals from two phases simultaneously. Experimental results demonstrate that the proposed method significantly outperforms traditional approaches, particularly in scenarios where the sample signals are noise-adding signals. Fault classification accuracy of the proposed improved convolutional neural network (MI-CNN) about 95.12% with noise-adding signals at the signal-to- noise ratio of 20 dB.
Volume: 15
Issue: 2
Page: 1656-1669
Publish at: 2025-04-01

Connected caregiving: investigating mothers in the era of digital access

10.11591/ijict.v14i1.pp347-354
Anissa Saidi , Wong Yee Von , Tirzah Zubeidah Zachariah@ Omar , Lim Seong Pek , Rita Wong Mee Mee , Khoo Kim Leng
Mothers have embraced and utilized digital access for nurturing and personal use to enhance their roles while balancing newfound demands. The Internet has provided mothers access to information on various topics, including pregnancy, childbirth, and infant care. Social media tools and platforms have also provided mothers with a space to connect with other mothers, share experiences, and seek support. This scoping review aims to identify the relationship of the focus skills among mothers in utilizing digital access. Four databases, including Scopus, web of science (WOS), education resources information centre (ERIC), and ScienceDirect, were used in this research, which found 36 articles for eligibility. Only 16 articles are eligible for analysis and reference after the exclusion and inclusion process for data collection. Based on the 16 publications examining mothers’ use of internet access, four essential skills have been identified. These included social, digital, cultural, and problem-solving skills and are acknowledged as being related to digital access mothering. The findings show these skills are offered to mothers through digital access, fostering diverse skill sets, contributing to their empowerment, and supporting sustainable development goal 5: gender equality, aiming to enhance women’s roles and ensure equal opportunities through digital inclusion.
Volume: 14
Issue: 1
Page: 347-354
Publish at: 2025-04-01

Two-scale decomposition and deep learning fusion for visible and infrared images

10.11591/ijece.v15i2.pp1593-1601
Ruhan Bevi Azad , Hari Unnikrishnan , Lokesh Gopinath
The paper focuses on the fusion of visible and infrared images to generate composite images that preserve both the thermal radiation information from the infrared spectrum and the detailed texture from the visible spectrum. The proposed approach combines traditional methods, such as two-scale decomposition, with deep learning techniques, specifically employing an autoencoder architecture. The source images are subjected to two-scale decomposition, which extracts high-frequency detail and low-frequency base information. Additionally, an algorithmic unravelling technique establishes a logical connection between deep neural networks and traditional signal processing algorithms. The model consists of two encoders for decomposition and a decoder after the unravelling operation. During testing, a fusion layer merges the decomposed feature maps, and the decoder generates the fused image. Evaluation metrics including entropy, average gradient, spatial frequency and standard deviation are employed to subjectively assess fusion quality. The proposed approach demonstrates promise for effectively combining visible and infrared imagery for various applications.
Volume: 15
Issue: 2
Page: 1593-1601
Publish at: 2025-04-01

Classification of brain stroke based on susceptibility-weighted imaging using machine learning

10.11591/ijece.v15i2.pp1602-1611
Shaarmila Kandaya , Norhashimah Mohd Saad , Abdul Rahim Abdullah , Ezreen Farina Shair , Ahmad Sobri Muda , Muhammad Izzat Ahmad Sabri
Magnetic resonance imaging (MRI) is used to identify brain disorders, particularly strokes. Rapid treatment, often referred to as "time is brain," is emphasized in recent studies, stressing the significance of early intervention within six hours of stroke onset to save lives and enhance outcomes. The traditional manual diagnosis of brain strokes by neuroradiologists is both subjective and time-intensive. To tackle this challenge, this study introduces an automated method for classify brain stroke from MRI images based on pre- and post-stroke patients. The technique employs machine learning, with a focus on susceptibility weighted imaging (SWI) sequences, and involves four stages: preprocessing, segmentation, feature extraction, classification and performance evaluation. The paper proposes classification and performance evaluation to determine stroke region according to three types of categories, those are poor improvement, moderate improvement and good improvement stroke patients based on pre and post patients. Then, performance evaluation is verified using accuracy, sensitivity and specificity. Results indicate that the hybrid support vector machine and bagged tree (SVMBT) yields the best performance for stroke lesion classification, achieving the highest accuracy which is 99% and showing significant improvement for stroke patients. In conclusion, the proposed stroke classification technique demonstrates promising potential for brain stroke diagnosis, offering an efficient and automated tool to assist medical professionals in timely and accurate assessments.
Volume: 15
Issue: 2
Page: 1602-1611
Publish at: 2025-04-01

Android-based smart digital marketplace application on agricultural commodities using a new variant recommendation system

10.11591/ijece.v15i2.pp1968-1977
Subiyanto Subiyanto , Sucihatiningsih Dian Wisika Prajanti , Nur Azis Salim , Setya Budi Arif Prabowo , Deyndrawan Sutrisno , Andika Anantyo , Dewi Anggriani
In the marketing of agricultural products, addressing the challenges associated with extensive distribution chains is essential, as these directly affect sellers. Additionally, the vast array of available product options often overwhelms customers, complicating their efforts to identify and purchase items that align with their preferences. This work aims to develop a smart e-commerce application for agribusiness, specifically designed for agricultural products on the Android platform. The application integrates a recommendation system that utilizes geolocation-aware neural graph collaborative filtering (GA-NGCF), which facilitates product marketing for farmers and streamlines the product search and selection process for users based on personalized preferences. The development process encompassed various stages, from planning to rigorous testing. The application’s recommendation system, which implements GA-NGCF, operates based on three primary elements: the creation of a geolocation graph of user-item data, the integration of information between neighboring nodes, and the prediction of user preferences. The resulting smart agribusiness e-commerce application, enhanced by GA-NGCF, demonstrated marked improvements in recommendation accuracy and overall application performance during testing. Empirical results indicated substantial enhancements in recommendation metrics, with GA-NGCF achieving a recall of 0.34, a precision of 0.36, and normalized discounted cumulative gain of 0.37, thereby outperforming existing models.
Volume: 15
Issue: 2
Page: 1968-1977
Publish at: 2025-04-01

Vehicle side control of a wireless power transfer charger using optimized artificial neural network

10.11591/ijece.v15i2.pp1487-1498
Marouane El Ancary , Abdellah Lassioui , Hassan El Fadil , Anwar Hasni , Yassine El Asri , Zakariae El Idrissi
This paper investigates a new approach to control a wireless power transfer (WPT) charger for electric vehicles (EVs) employing an optimized artificial neural network (ANN). Enhancing the efficiency and robustness of such systems is crucial, and integrating artificial intelligence (AI)-based solutions has introduced innovative approaches in this field. The proposed method enables precise regulation of battery charging voltage even under challenging conditions, such as coil misalignment or shared grounding assemblies for multiple EVs. To assess the stability and robustness of the proposed controller, its performance was evaluated under scenarios of coil misalignment and shared grounding assemblies for EVs with varying battery voltages. The controller effectively eliminated overshoot and significantly reduced residual output voltage ripple by 4.33% compared to a conventional proportional-integral (PI) controller, demonstrating the superior performance and reliability of the ANN-based control approach.
Volume: 15
Issue: 2
Page: 1487-1498
Publish at: 2025-04-01

Enhancing engineering education through virtual reality: a systematic study on immersive engineering education practices

10.11591/ijece.v15i2.pp1889-1899
Tarek Riaji , Sanae El Hassani , Fatima Ezzahrae M'hamdi Alaoui
This article explores the integration of virtual reality (VR) and associated technologies in engineering education, focusing on the pedagogical approaches adopted in this integration, which we refer to as immersive engineering education. This study considers the application possibilities and the transformative impact of VR on engineering education. The article addresses the critical collection and analysis of VR applications in engineering education. It covers main VR-related papers published from 2015 to February 2024 and indexed in Scopus, Web of Sciences, or both, and discussing design, development challenges, and collaborative tools. Empirical evidence showcases improved engagement, motivation, and learning outcomes. The findings offer modern insights for educators and researchers on leveraging VR for impactful learning experiences, while also noting the need for further research in this evolving field.
Volume: 15
Issue: 2
Page: 1889-1899
Publish at: 2025-04-01

An improved key scheduling for advanced encryption standard with expanded round constants and non-linear property of cubic polynomials

10.11591/ijece.v15i2.pp2455-2467
Muthu Meenakshi Ganesan , Sabeen Selvaraj
The advanced encryption standard (AES) offers strong symmetric key encryption, ensuring data security in cloud computing environments during transmission and storage. However, its key scheduling algorithm is known to have flaws, including vulnerabilities to related-key attacks, inadequate nonlinearity, less complicated key expansion, and possible side-channel attack susceptibilities. This study aims to strengthen the independence among round keys generated by the key expansion process of AES—that is, the value of one round key does not reveal anything about the value of another round key—by improving the key scheduling process. Data sets of random, low, and high-density initial secret keys were used to evaluate the strength of the improved key scheduling algorithm through the National Institute of Standards and Technology (NIST) frequency test, the avalanche effect, and the Hamming distance between two consecutive round keys. A related-key analysis was performed to assess the robustness of the proposed key scheduling algorithm, revealing improved resistance to key-related cryptanalysis.
Volume: 15
Issue: 2
Page: 2455-2467
Publish at: 2025-04-01
1 2 3
...
2097 2098
Show 327 of 2098

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