cover
Contact Name
Ramdan Satra
Contact Email
Ramdan Satra
Phone
-
Journal Mail Official
ramdan@umi.ac.id
Editorial Address
-
Location
Kota makassar,
Sulawesi selatan
INDONESIA
ILKOM Jurnal Ilmiah
ISSN : 20871716     EISSN : 25487779     DOI : -
Core Subject : Science,
ILKOM Jurnal Ilmiah is an Indonesian scientific journal published by the Department of Information Technology, Faculty of Computer Science, Universitas Muslim Indonesia. ILKOM Jurnal Ilmiah covers all aspects of the latest outstanding research and developments in the field of Computer science, including Artificial intelligence, Computer architecture and engineering, Computer performance analysis, Computer graphics and visualization, Computer security and cryptography, Computational science, Computer networks, Concurrent, parallel and distributed systems, Databases, Human-computer interaction, Embedded system, and Software engineering.
Arjuna Subject : -
Articles 617 Documents
Development of an Intelligent Catch the Stick System for Measuring Human Motor Coordination and Reaction Speed Wirawan, Nanda Tommy; Ernes, Risa Nadia
ILKOM Jurnal Ilmiah Vol 18, No 1 (2026)
Publisher : Prodi Teknik Informatika FIK Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/ilkom.v18i1.2887.126-137

Abstract

Conventional clinical methods for assessing sensorimotor coordination, such as the Fugl Meyer Assessment (FMA) and Action Research Arm Test (ARAT), often lack the objectivity and high-resolution sensitivity required to detect subtle micro improvements in motor performance. This study presents the design, development, and validation of an intelligent "Catch the Stick" system aimed at accurately and quantitatively assessing human sensorimotor coordination and reaction speed. The proposed multi-metric system integrates 9 axis inertial measurement units (IMUs), a 60 fps computer vision tracking system, and algorithmic classification to evaluate real-time temporal and spatial responses during random stick-dropping tasks. An experimental study was conducted involving fifteen participants (10 healthy individuals and 5 clinical patients with mild to moderate sensorimotor deficits) tested under varying stimulus loads ranging from 1 to 10 sticks. The system demonstrated strong to excellent test-retest reliability (ICC 0.75) and high detection precision (±15 ms temporal error, 2 mm spatial error). Experimental results revealed that increased stick quantity directly correlated with prolonged reaction times, thereby objectively quantifying cognitive motor overload. Furthermore, the system exhibited strong concurrent validity with conventional tools, showing significant positive correlations with FMA (r = 0.78) and ARAT (r = 0.74) scores. Notably, the intelligent system proved more sensitive to micro improvements in 72% of participants compared to traditional clinical scales, although ceiling effects were observed in low difficulty tasks among healthy users. Overall, the intelligent Catch the Stick platform offers a robust, scalable, and highly sensitive solution for quantifying sensorimotor performance in clinical settings, laying the foundation for future robotic automation and autonomous training protocols
Implementation Of Deep Learning Using Convolutional Neural Network Method In A Rupiah Banknote Detection System For Those With Low Vision Akhiyar, Dinul; Tukino, Tukino; Defit, Sarjon
ILKOM Jurnal Ilmiah Vol 17, No 1 (2025)
Publisher : Prodi Teknik Informatika FIK Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/ilkom.v17i1.2253.34-43

Abstract

The application of deep learning in various sectors continues to grow due to its ability to provide efficient and effective solutions to complex problems. One significant implementation is in object detection, such as identifying Indonesian rupiah banknotes. This innovation aims to assist individuals with visual impairments in using money more effectively. At present, visually impaired individuals rely on conventional methods, such as identifying banknotes by touch, folding them in specific ways, or seeking assistance from others. However, these methods are often time-consuming, prone to error, and lack practicality in everyday situations. In this project, a system was developed using the Convolutional Neural Network (CNN) architecture combined with the YOLO (You Only Look Once) algorithm. YOLO is renowned for its speed and accuracy in real-time object detection, making it an ideal choice for detecting banknotes in moving images. The training dataset included 1,260 images, and the model underwent 7,000 iterations during training. As a result, the system achieved a high mean Average Precision (mAP) score of 97.65%, demonstrating its robustness and precision. For validation, 140 test images were utilized, which yielded an impressive mAP value of 97.5%. To further evaluate the system's reliability, tests were conducted under varying conditions, such as banknotes with creases, folds, or different lighting scenarios. These tests resulted in an mAP score of 88%, showcasing the system's adaptability to real-world conditions. This system provides significant benefits for individuals with visual impairments by offering a practical, efficient, and accurate solution for recognizing banknotes. With this technology, visually impaired users can interact with currency independently, reducing their reliance on others and traditional, less practical methods. This innovation not only enhances their autonomy but also fosters inclusivity in financial transactions. By integrating this system into mobile applications or wearable devices, its accessibility and usability can be further improved, paving the way for a broader societal impact.
Enhanced Violence Detection in CCTV Using LSTM Hasanudin, Muhaimin; Santoso, Hadi; Wahab, Abdi; Indrianto, Indrianto; Kuswardani, Dwina; Ridlan, Ahmad
ILKOM Jurnal Ilmiah Vol 17, No 2 (2025)
Publisher : Prodi Teknik Informatika FIK Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/ilkom.v17i2.2318.196-202

Abstract

Violence detection in CCTV footage remains a critical challenge for public safety, necessitating automated solutions to overcome human monitoring limitations. This study proposes an LSTM-based framework to improve detection accuracy by analyzing temporal patterns in surveillance videos. Using a dataset of 2,000 videos (1,000 violent/1,000 non-violent), the model extracts spatial-temporal features via optical flow and achieves 93% training accuracy and 91% test accuracy, with a precision of 92% and AUC of 0.94. Results demonstrate significant improvements over traditional methods, particularly in dynamic scenarios, though performance dips for occluded actions or weapon-related violence. The discussion highlights the model’s real-time applicability, computational efficiency (120 ms latency per segment), and alignment with smart city surveillance needs. Limitations include dataset diversity and environmental variability, suggesting future directions in multi-modal data fusion and edge computing. This research advances AI-powered security systems, offering a robust tool for proactive threat detection while underscoring the need for scalable, context-aware solutions.
Isolation Forest-Based Anomaly Detection in IoT Smart Home Network Traffic Luthfi, Ahmad; Emigawaty, Emigawaty
ILKOM Jurnal Ilmiah Vol 18, No 1 (2026)
Publisher : Prodi Teknik Informatika FIK Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/ilkom.v18i1.3156.43-57

Abstract

The convergence of the Internet of Things (IoT) and Society 5.0 has successfully led to a human-centered and data-driven life ecosystem. IoT has become the backbone for infrastructure implemented in various domains, ranging from smart homes and smart farming to smart industrial environments. Nevertheless, as IoT devices become more connected and integrated into the ecosystem, the attack surface expands and network security becomes more challenging. The massive convergence and connectivity of IoT devices have a high potential for attacks on network infrastructure, such as Denial of Service (DoS), port scanning, exfiltration, brute force, and man-in-the-middle attacks. This study aims to detect anomalies in IoT network traffic by applying the Isolation Forest (IF) algorithm. The dataset was obtained from an IoT gateway connected to smart home devices and includes features such as data packet size, connection duration, source and destination capacity, attack protocols used, and the connection status of each device. The experimental results of this study indicate that the IF method can identify smart home device attacks with a competitive level of accuracy. The results of the anomaly analysis are then presented through a confusion matrix, classification report, and analytical visualizations such as 2D PCA, t-SNE, heatmap, and temporal distribution of anomalies. This study declares that the IF method contributes effectively to the analysis of Intrusion Detection Systems (IDS) in IoT environments such as smart homes that are heterogeneous and dynamic
Sentiment Analysis towards Jokowi Post-Presidential Term Using CNN-BiLSTM with Multi-head Attention on Platform X Setyawan, Muhammad Rizki; Putra, Fajar Rahardika Bahari; Ramadhani, Ardhina
ILKOM Jurnal Ilmiah Vol 17, No 2 (2025)
Publisher : Prodi Teknik Informatika FIK Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/ilkom.v17i2.2843.150-161

Abstract

The development of social media has changed the way the public expresses political opinions, especially regarding the evaluation of President Joko Widodo’s (Jokowi) leadership after his term. Platform X (formerly Twitter) has become the primary source of public opinion data, but the use of informal language and sarcasm makes accurate sentiment analysis challenging. This study creates a sentiment analysis model that uses deep learning with a CNN-BiLSTM structure and a multi-head attention mechanism. The dataset consists of 52,643 tweets that have been labeled and embedded using IndoBERT. To address class imbalance, the SMOTE method was applied to the training data, enabling the model to better learn from minority classes. The results indicate that the model achieves a high accuracy of 98.78%, with an average precision, recall, and F1-score of 0.98. These findings indicate that the model is not only accurate but also reliable in distinguishing each sentiment class. A comparison with other model variants suggests that the complete combination of CNN-BiLSTM and Multi-Head Attention delivers the best performance, although the improvement is relatively small.
K-Means and K-Medoid in Clustering Analysis of Network Congestion Level Darwis, Herdianti; Purnawansyah, Purnawansyah; Umalekhoa, Alfi Syahrin; Adnan, Adam; Salim, Yulita; Umar, Fitriyani; Raja, Roesman Ridwan; Fajar AR, Muh. Aqil
ILKOM Jurnal Ilmiah Vol 17, No 3 (2025)
Publisher : Prodi Teknik Informatika FIK Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/ilkom.v17i3.2083.323-335

Abstract

This research investigates the application of clustering techniques to network congestion data at Universitas Muslim Indonesia, employing a hybrid metric approach based on packet loss and delay. The study utilized two algorithms, K-Means and K-Medoid, applied in a semi-supervised scenario to group 255,147 network data points into 3, 4, and 5 clusters, considering 10 principal variables. During the pre-processing phase, data cleansing was conducted to address missing values, followed by normalization to standardize the scale of numerical variables, thereby preparing the data for the clustering process. Model validation was performed using four cluster evaluation methods: Gap Statistic, Davies-Bouldin Index, and Elbow Method. The evaluation results indicate that both algorithms were capable of forming valid and reliable clusters. However, the K-Means algorithm demonstrated superior performance compared to K-Medoid, particularly when utilizing three Quality of Service variables: throughput, packet loss, and delay. In this configuration, K-Means yielded more stable clusters, a clearer separation between clusters, and a more structured visualization. Consequently, K-Means is considered more optimal for classifying network congestion levels and presents an effective approach for network data segmentation
SqueezeNet Image Embedding and Support Vector Machine for Recognizing Hand Gestures in Indonesian Sign Language System Islami, Megan Shahra; Jamzuri, Eko Rudiawan
ILKOM Jurnal Ilmiah Vol 17, No 2 (2025)
Publisher : Prodi Teknik Informatika FIK Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/ilkom.v17i2.2476.98-106

Abstract

This research proposes a hand gesture recognition method for the Sistem Isyarat Bahasa Indonesia (SIBI) sign language, integrating SqueezeNet for image feature extraction and Support Vector Machine (SVM) for classification. The study focuses on 24 static gestures representing alphabetic letters, excluding J and Z due to their motion-based representation. The dataset consists of 5280 RGB images (227×227 pixels), with 220 samples per gesture, obtained from a public Kaggle source. SqueezeNet, a lightweight CNN architecture, is used to generate 1000-dimensional feature vectors, which are then classified using an SVM with an RBF kernel (C = 1.0) to effectively handle non-linear decision boundaries. A 10-fold cross-validation was applied without data augmentation to evaluate baseline performance. The proposed method achieved 99.51% classification accuracy, with an average precision of 94.04%, recall of 94.02%, and F1-score of 94.02%. Certain gestures, such as G, H, and Q, achieved near-perfect recognition, while others, like V, presented greater classification challenges with a recall of 80.5%. Compared to existing models such as MobileNet (98% accuracy) and VGG16 (86% accuracy) on the same dataset, the SqueezeNet–SVM combination provides competitive or superior accuracy with significantly reduced computational requirements. These results highlight the method’s potential for real-time integration into mobile or embedded sign language translation applications, bridging communication gaps between the deaf and hearing communities. Future work will focus on improving performance for difficult gestures, applying data augmentation to enhance generalization, and developing a prototype mobile application for real-world testing in relevant environments.
An Exploration of the Work Performance of Educators in Transformative Schools: Leveraging Machine Learning for Performance Insights Maulidi, Rakhmad; Palandi, Jozua Ferjanus; Kristanto, Bagus Kristomoyo; Isyriyah, Laila; Rahmatullah, Rizky; Adi, Puput Dani Prasetyo; Kitagawa, Akio
ILKOM Jurnal Ilmiah Vol 18, No 1 (2026)
Publisher : Prodi Teknik Informatika FIK Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/ilkom.v18i1.2358.109-125

Abstract

Education has gone through various phases, and entered the transformative school mode which can be said to change the existing order of the previous schooling system or procedures, because many modes can be done in the transformative school, students can learn in school buildings or classes, or in the field or real industry or the real world of work, with the introduction of a wider and more complex world, this is one of them. This research tries to create and analyze transformative schools in 3 algorithms, namely regression algorithms, classification algorithms, and clustering algorithms that provide a detailed analysis of the results of the analysis of transformative schools currently promoted by the government. from the results of the analysis raises performance conclusions, and in this phase a conclusion can be drawn whether the Transformative school is able to provide answers about the performance of teachers, students, teacher education levels, school locations, number of students, learning methods, or any paramaters that can provide detailed and detailed answers to get performance analysis from Machine Learning, and Work Performance of teachers in Transformative schools with precision. Quantitatively, the value of performance is determined by innovation by 43.2%, followed by technological capabilities and collaboration, 27.9% and 17.2% respectively. and based on cluster level, cluster 3 is the best with 118 educators, cluster 0, 127 educators with high innovators, and cluster 2, 126 educators, and cluster 1 with 129 educators. and from the paradox of transformative practices 30.6% are high Adopters
Refining the Performance of Neural Networks with Simple Architectures for Indonesian Sign Language System (SIBI) Letter Recognition Using Keypoint Detection Amir, Nur Hikma; Dewa, Chandra Kusuma; Luthfi, Ahmad
ILKOM Jurnal Ilmiah Vol 17, No 1 (2025)
Publisher : Prodi Teknik Informatika FIK Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/ilkom.v17i1.2522.64-73

Abstract

The diversity of non-verbal communication styles among persons with disabilities in Indonesia highlights the urgent need for technological solutions that support accessibility in both workplace settings and social contexts. This study proposes a novel approach to improving neural network performance through the use of simple architectures for recognizing Indonesian Sign Language (SIBI) letters M and N, by applying keypoint detection while accounting for hand size variations (17–22 cm). Four models were evaluated: YOLOv5 based on image detection, as well as VGG-16, Attention, and Multi-Layer Perceptron (MLP) developed using keypoint detection. The evaluation was conducted in real-time, taking into account accessories such as rings, watches, and gloves, as well as varying lighting intensities to simulate real-world user environments. The novelty lies in the integration of keypoint detection into lightweight architectures, which significantly improves accuracy and resilience against visual disturbances (noise). The MLP model achieved the best performance, with an accuracy of 94% for M and 93% for N, outperforming more complex approaches such as YOLOv5, which showed a significant drop in accuracy under disturbed conditions. The integration of VGG-16 with Attention resulted in underfitting, emphasizing that complexity does not always correlate with effectiveness. These findings underscore the potential of lightweight models to enhance technological accessibility for the disabled community across various social and professional domains.
Evaluating the Effectiveness of TBaWI for Imputation of Missing Rainfall Data Syafie, Lukman; Awangga, Narendra; Salim, Yulita
ILKOM Jurnal Ilmiah Vol 18, No 1 (2026)
Publisher : Prodi Teknik Informatika FIK Universitas Muslim Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33096/ilkom.v18i1.3273.97-108

Abstract

Daily rainfall data plays an important role in hydrological and climatological analysis, especially in tropical regions characterised by high rainfall variability and sharp seasonal changes. However, observational data often has gaps, which can reduce model accuracy and obscure relevant climatological signals. This study addresses these issues by applying the Trend-Based Adaptive Window Imputation (TBaWI) method, an adaptive imputation approach that considers local temporal trends and seasonal dynamics in estimating missing rainfall values. This method was tested using CHIRPS data for the Makassar region for the period 2014–2023 with synthetic data loss scenarios of 10%, 15%, 20%, and 25%. The results show that TBaWI consistently provides a lower Mean Absolute Error (MAE) value, namely 6.14–7.65 mm, compared to linear interpolation, which produces 6.46–7.75 mm. The SMAPE value of TBaWI is also lower, for example 33.16% in the 15% data loss scenario, compared to interpolation at 35.06%. In addition, this method showed an improvement in the ability to identify dry days through the Zero Hit Rate (ZHR), which reached 60.08% in the 20% data loss scenario, higher than the interpolation of 58.32%, while the Rainy Hit Rate (RHR) remained in a stable range of 79–88%. These findings indicate that TBaWI is more effective in maintaining climatological consistency and numerical accuracy of tropical rainfall data. Further research is expected to integrate spatial aspects and optimise machine learning-based parameters to improve the generalisation of the method under various climatic conditions.