Indonesian Journal of Electrical Engineering and Informatics (IJEEI)
Indonesian Journal of Electrical Engineering and Informatics (IJEEI) is a peer reviewed International Journal in English published four issues per year (March, June, September and December). The aim of Indonesian Journal of Electrical Engineering and Informatics (IJEEI) is to publish high-quality articles dedicated to all aspects of the latest outstanding developments in the field of electrical engineering. Its scope encompasses the engineering of Telecommunication and Information Technology, Applied Computing & Computer, Instrumentation & Control, Electrical (Power), Electronics, and Informatics.
Articles
825 Documents
<![CDATA[Enhancing Recommendation Systems via Aspect-Level Sentiment Analysis for Identifying Unreliable Reviews ]]>
<![CDATA[El Mehdi Lghaouch]]>;
<![CDATA[Fatima Zahra Abbour]]>;
<![CDATA[Soumaya Ounacer]]>;
<![CDATA[Soufiane Ardchir]]>;
<![CDATA[Mohamed Azzouazi]]>
<![CDATA[ Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol 14, No 2: June 2026 ]]>
Publisher : <![CDATA[IAES Indonesian Section ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.52549/ijeei.v14i2.7834
<![CDATA[User-generated reviews are an important source of information for e-commerce recommendation systems. However, reviews may contain conflicting opinions across product aspects such as quality, price, and delivery, which can weaken both review-level trust and aspect-level signals. To address this issue, we propose AspectGuard, a filtering framework based on Aspect-Based Sentiment Analysis (ABSA) and a Sentiment Divergence Score (SDS) that quantifies disagreement across aspect-level sentiments within a review. Reviews with high divergence are identified as potentially unreliable and excluded before constructing recommendation profiles. To isolate the contribution of the proposed filtering mechanism, we compare the same content-based recommendation model with and without AspectGuard filtering. Experiments conducted on a subset of the Amazon Reviews 2023 Electronics dataset show that the proposed filtering strategy improves both rating prediction and ranking quality. These results indicate that aspect level sentiment divergence can serve as an effective signal for identifying unreliable reviews and improving the reliability of recommendation systems.]]>
<![CDATA[Dam Hazard Prediction Using an AIoT Based System ]]>
<![CDATA[Nabila Aissani]]>;
<![CDATA[Abderraouf Messai]]>;
<![CDATA[Salheddine Sadouni]]>
<![CDATA[ Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol 14, No 2: June 2026 ]]>
Publisher : <![CDATA[IAES Indonesian Section ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.52549/ijeei.v14i2.7376
<![CDATA[Dams are a very important infrastructure because they are one of the main water and electricity supplies for most countries; However, it is susceptible to demolition due to several factors such as aging and floods. This not only causes damage to water supplies and economical loss, but also human lives. And with the current climate change and global warming, heavy rainfalls and storms leading to floods are rapid and abrupt; existing studies tend to focus on one aspect rather than a total security system. Which drives us to look for the answer to the most important question; How can we deploy a system that is capable predicting coming floods and estimating the impact giving us enough time for a reaction? Hence, in this paper, we propose an Artificial intelligence of things based system to predict potential floods in dams by overtopping and potential breaching, their time, and estimated threatened distance. This system includes mainly 1) several types of sensors like temperature, humidity, and water level 2) AI models like one class support vector machine, isolation forest, hybrid models, and regression models. The results were promising for improving dam security.]]>
<![CDATA[Boosting Few-Shot Text Classification in Large Language Models with Data Augmentation ]]>
<![CDATA[Ahmed El Saeid Ali Soliman]]>;
<![CDATA[Reda Abd Elwahab El-Khoribi]]>;
<![CDATA[Basma El-Demerdash]]>;
<![CDATA[Ahmed Elgayar]]>
<![CDATA[ Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol 14, No 2: June 2026 ]]>
Publisher : <![CDATA[IAES Indonesian Section ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.52549/ijeei.v14i2.7074
<![CDATA[Few-shot text classification is a challenging problem in natural language processing. Models have to generalize from a small number of labeled examples. This paper investigates the effectiveness of data augmentation methods (back-translation, paraphrasing, and noise injection) on overfitting prevention and generalization enhancement in few-shot scenario for small transformer models such as DistilBERT and DistilRoBERTa. We perform experiments on two standard benchmark datasets, DBPedia-14 and BBC News, in the 3-shot, 5-shot, and 7-shot settings. The result shows that the data augmentation substantially improves the performance of the classification. With back-translation, DistilBERT achieves 0.98 and 0.96 accuracy on DBPedia-14 and BBC News in 7-shot setting, compared to 0.88 and 0.86 accuracy without any augmentation. These results demonstrate that carefully selected augmentations can bridge the performance gap between few-shot and fully supervised learning, enabling competitive results on resource constrained hardware without the need for massive, labeled datasets.]]>
<![CDATA[Lightweight Enhancement Module for Efficient Sea Turtle Species Detection under Illumination Variation ]]>
<![CDATA[Yuliana Mose]]>;
<![CDATA[Alex Copernikus Andaria]]>;
<![CDATA[Hebron Prasetya]]>;
<![CDATA[Muhamad Dwisnanto Putro]]>
<![CDATA[ Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol 14, No 2: June 2026 ]]>
Publisher : <![CDATA[IAES Indonesian Section ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.52549/ijeei.v14i2.7399
<![CDATA[Sea turtle monitoring is crucial for maintaining ecosystem balance, although traditional methods are costly and inefficient. Object detection provides a foundation for autonomous monitoring by localizing sea turtles in their natural habitat. At the same time, recent deep learning models can operate in real-time. On the other hand, underwater detection remains challenging due to extreme illumination, underwater blur, and underwater distortions, which obscure key features from the background. To address these issues, this work proposes a new sea turtle detector, based on a modified YOLOv12-nano model that incorporates a lightweight enhancement module to improve robustness under diverse illumination. The network introduced Efficient Spatial Convolution (ESC) at the end of the backbone to improve the representation of high-level features. It involves Dual Spatial Convolution (DSC) to improve backbone feature extraction with a large kernel size while preserving efficiency. The Lite module is designed to enrich feature diversity and selectively enhance the quality of high-level features. These modifications significantly enhance detection performance under diverse underwater lighting conditions. The proposed network achieves higher precision than the original YOLOv12-nano while maintaining efficiency, making it well-suited for deployment on low-cost devices.]]>
<![CDATA[Implementation and Validation of the Optimized Space Diversified Antenna Arrays for Extended FM-RF Propagation ]]>
<![CDATA[Fernando Nuena Mangubat Jr.]]>;
<![CDATA[Naomi Asendiente Bajao]]>;
<![CDATA[Victorino Hermosilla Patindol]]>;
<![CDATA[Florieza Mendoza Mangubat]]>
<![CDATA[ Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol 14, No 2: June 2026 ]]>
Publisher : <![CDATA[IAES Indonesian Section ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.52549/ijeei.v14i2.7649
<![CDATA[FM broadcast antennas come in many shapes and sizes, but the dipole remains one of the most widely used designs because of its simple structure and reliable performance. For long-distance FM transmission, multiple antennas are often stacked to improve coverage and produce a stronger, more consistent radiation pattern. In these configurations, each antenna bay is adjusted to achieve the desired directional characteristics, and identical bays generate the same azimuth radiation pattern when installed as an array. The researchers designed and optimized a low-power FM broadcast antenna that performed better than the conventional design while meeting the required polarization and radiation pattern specifications based on antenna simulation results. However, because of the restrictions brought about by the COVID-19 pandemic, the antenna was never fabricated or tested using an antenna analyzer. As a result, its performance had only been evaluated through simulation. This study bridges that gap by fabricating the proposed antenna and validating its performance under actual operating conditions. The measured parameters obtained from an antenna analyzer were compared with the simulation results, including measurements taken after the antenna was mounted on a steel tower to determine the effects of the installation environment. The results showed close agreement between the simulated and measured values. The fabricated antenna produced a low reflection coefficient and a voltage standing wave ratio (VSWR) close to 1.0, indicating efficient power transfer and good impedance matching. Field measurements using the Deva Radio Explorer-Mobile FM Radio Analyzer also confirmed stronger signal performance than the conventional antenna. These findings demonstrate that the proposed design performs well not only in simulation but also in practical deployment, making it a viable option for low-power FM broadcasting.]]>
<![CDATA[Enhancing Arabic Diacritization Using BERT with BiLSTM and BiGRU Heads ]]>
<![CDATA[Ahmed Abdelhamed Elgayar]]>;
<![CDATA[Abdul-Hadi Nabi Ahmed]]>;
<![CDATA[Basma Ezzat El-Demerdash]]>;
<![CDATA[Ahmed E. Soliman]]>
<![CDATA[ Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol 14, No 2: June 2026 ]]>
Publisher : <![CDATA[IAES Indonesian Section ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.52549/ijeei.v14i2.7033
<![CDATA[Arabic diacritization is the task of adding diacritical marks to Arabic text to aid correct articulation, resolve lexical ambiguity, and support downstream natural language processing tasks. This study proposes a hybrid deep learning architecture that augments a pre-trained transformer encoder (AraBERTv02) with a sequential prediction head built on either a Bidirectional Long Short-Term Memory (BiLSTM) or a Bidirectional Gated Recurrent Unit (BiGRU) network, replacing the conventional linear classification layer. This design introduces an explicit sequential inductive bias at the output level, enabling the model to capture local label dependencies that are not modeled by a standard linear head. The models were trained and evaluated on the Tashkeela Processed (TP) benchmark corpus of fully diacritized Modern Standard Arabic. The BiLSTM based model achieves a Diacritic Error Rate (DER) of 1.82% and a Word Error Rate (WER) of 2.58% while the BiGRU variant has a DER of 1.84% and a WER of 2.62% both outperforming the prior state-of-the-art on the same benchmark (WER: 3.34%). Error analysis reveals a systematic confusion of short vowels, especially in verb-initial contexts where fatha and damma alternate to mark active and passive voice, illustrating the inherent morphological complexity of Arabic diacritization. The proposed architecture is also shown to be suitable for real-world deployment with an inference time of 9.22 ms and 8.74 ms per sentence for BiLSTM and BiGRU models respectively. Thus, both variants are compatible with near real-time applications such as Arabic text-to-speech systems and educational tools.]]>
<![CDATA[TurtleNet: Explainable Turtle Species Classification Using Attention-Enhanced EfficientNetB4 and Ensemble Learning ]]>
<![CDATA[Pragyat Jyoti Baruah]]>;
<![CDATA[Bugge Prince Solomon Raj Kumar]]>;
<![CDATA[Arun Jyoti Nath]]>;
<![CDATA[Arnab Paul]]>;
<![CDATA[Jarine Anjum]]>;
<![CDATA[Akash Nath]]>;
<![CDATA[Tirthanka Borah]]>
<![CDATA[ Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol 14, No 2: June 2026 ]]>
Publisher : <![CDATA[IAES Indonesian Section ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.52549/ijeei.v14i2.7994
<![CDATA[Turtles are remarkable organisms as they play a vital role in maintaining a balanced and healthy ecosystem. However, their existence is at risk due to growing threats of habitat loss, illegal trade and environmental degradation. To support conservation and monitoring efforts for these turtle species, this study proposes TurtleNet, a custom deep learning model designed to accurately identify turtle species with acute precision. The proposed model relies on a customized EfficientNetB4 architecture as a backbone with an enhanced lightweight attention mechanism of Squeeze and Excitation blocks, allowing it to focus on subtle visual features that distinguish closely related species. The model was trained and evaluated on our curated Turtle1 dataset as well as the images.cv turtle dataset. The performance comparison of TurtleNet with 11 individual state-of-the-art deep learning classification models showed its superior ability by classifying Asian turtle species with 97.25% accuracy. To further enhance classification performance, TurtleNet is integrated with the YOLOv8 model using a soft-voting ensemble strategy, thereby increasing the classification accuracy by 2.41%. The results demonstrate that TurtleNet has strong potential for automated species monitoring and conservation using non-invasive measures. Grad-CAM visualization is used to highlight discriminative morphological regions relevant to image identification and is used to interpret the feature selection process of the proposed model.]]>
<![CDATA[A Formal Methodology for Workload-Driven Comparative Evaluation of Cloud Infrastructure Architectures ]]>
<![CDATA[Valeriy Kozlovskiy]]>;
<![CDATA[Yurii Lysetskyi]]>;
<![CDATA[Nikita Bakun]]>;
<![CDATA[Diana Kozlovska]]>;
<![CDATA[Nataliia Yakymchuk]]>
<![CDATA[ Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol 14, No 2: June 2026 ]]>
Publisher : <![CDATA[IAES Indonesian Section ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.52549/ijeei.v14i2.7909
<![CDATA[In this paper we propose a formal approach for workload-driven comparative evaluation of cloud infrastructure architectures that support objective architectural decision making in dynamic operating scenarios. The novelty of the work is an integrated framework, which combines cost modeling that is workload dependent, autoscaling behavior, dynamic pricing models and infrastructure as code (IaC)-based representations of architectures in one single evaluation methodology. Different to former approaches which employ pricing calculation tools that fix the prices, or report individual cost measures, or use expertise of humans, the proposed approach represents architectural candidates as scenario-dependent systems, which are able to provide accurate cost and performance estimation based on the intensity of workload and scaling dynamics. Real architectures instead of just infrastructure models can be evaluated by formally representing an IaC configuration in terms of the used analytic cost and performance models. We evaluate our methodology using production data of a real-world application and prove our model has a 94% accuracy in estimating costs and significantly reduce architectural design and evaluation time.]]>
<![CDATA[Systematic Approach to NoSQL Indexing: Validating the ESR (Equality, Sort, Range) Heuristic for Query Performance ]]>
<![CDATA[Narut Butploy]]>;
<![CDATA[Kanokwan Khiewwan]]>;
<![CDATA[Jindaporn Ongate]]>;
<![CDATA[Thep Kueathaweekun]]>;
<![CDATA[Pakin Maneechot]]>
<![CDATA[ Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol 14, No 2: June 2026 ]]>
Publisher : <![CDATA[IAES Indonesian Section ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.52549/ijeei.v14i2.7670
<![CDATA[In this paper we describe the ESR heuristic, a simple and repeatable way to build compound indexes in MongoDB. The method prioritizes the query requirements in order of equality, sort, and range. The ESR heuristic can boost the query performance without the need for a complicated machine learning model. The experiment was conducted on four real-world datasets and two kinds of mixed queries on Equality, Sort, and Range (ESR). The results show that the systematic technique for compound indexing based on the ESR principle reduces the mean query latency in most cases. For example, the Online Retail dataset had a clear improvement with p-values less than 0.01 in both Welch’s t-test and the Mann-Whitney U-test. For the NYC Taxi dataset, the non-parametric test indicated a significant change in the latency distribution (p=0.032), implying that many queries have decreased tail latency. The Abalone dataset indicated a favorable trend, but it was not very significant statistically . The filter condition had low selectivity, hence only a little improvement was observed in the Bank Marketing dataset. The experiment also highlighted the COLLSCAN contradiction where MongoDB still scanned the entire collection with a proper compound index. However, the ESR heuristic decreased variance and tail latency in most circumstances.]]>
<![CDATA[A Pseudo-spectral Method with Least Squares Approach and Monomial Basis Functions for Fractional Ordinary Differential Equations ]]>
<![CDATA[Olumuyiwa Otegbeye]]>;
<![CDATA[Nancy Mukwevho]]>;
<![CDATA[Shina Daniel Oloniiju]]>
<![CDATA[ Indonesian Journal of Electrical Engineering and Informatics (IJEEI) Vol 14, No 2: June 2026 ]]>
Publisher : <![CDATA[IAES Indonesian Section ]]>
Show Abstract
|
Download Original
|
Original Source
|
Check in Google Scholar
|
DOI: 10.52549/ijeei.v14i2.6193
<![CDATA[The present work combines the pseudo-spectral method with least square technique to approximate non-linear and linear differential equations that are bounded and it makes use of monomials as trial functions. The pseudo-spectral method involves the approximation of solutions to differential equations by taking the sum of truncated basis functions, and this feature differentiates this method from difference-based methods. The idea of the least square technique is to optimally minimize the sum of squares of the residual function of the differential equation. The Caputo fractional differential operator is generally used given that it allows for the inclusion of the conditions at the initial state and the boundaries in formulating the equation(s). Non-linear fractional differential equations are linearized using the quasilinearization approach, which is based on the Newton-Raphson method. The resulting linear algebraic equations are regularized using the Tikhonov regularization technique due to the density of the equations. Some problems are solved to demonstrate the efficiency of the proposed method and display its convergence and accuracy, respectively, of approximate solutions obtained. Comparison of the numerical solutions and exact solutions (where obtainable) is conducted, and the numerical results are further presented in tabular form and graphically. The results obtained show that the proposed method is computationally efficient. The process of regularizing the numerical schemes of non-linear fractional differential equations was also shown to give more accurate solutions.]]>
