Publication Search

This study presents a Graph-Augmented Isolation Forest (GAIF), an unsupervised anomaly-detection framework for analyzing mobile user behavior. The proposed framework represents users and behavioral attributes as a user–feature bipartite graph, enabling the capture of relational dependencies that are not explicitly modeled in conventional vector-based approaches. Low-dimensional user representations are learned through Node2Vec and Graph Sample and Aggregate (GraphSAGE), and the resulting embeddings are subsequently processed by an Isolation Forest to produce anomaly scores. Experiments are conducted on a Mobile Device Usage and User Behavior dataset comprising 700 user profiles derived from application-level behavioral indicators. The dataset is treated as a behavioral abstraction rather than as a malware classification benchmark. A consistent 80:20 stratified train–test split is employed, with all learning-capable operations restricted to the training data to mitigate information leakage. Detection performance is evaluated post hoc using precision, recall, F1-score, and area under the curve (AUC) metrics. Under the evaluated setting, GAIF achieves an F1-score of 0.94 and an AUC of 0.97, demonstrating improved anomaly detection effectiveness relative to representative unsupervised baseline methods. These results are obtained on a static, proxy dataset and should not be interpreted as evidence of real-time deployment capability. Model interpretability is supported through post-hoc Uniform Manifold Approximation and Projection (UMAP) visualizations of the learned embeddings, providing structural insights into anomalous user behavior. Overall, the findings indicate that integrating graph-based representation learning with isolation-based anomaly scoring constitutes a computationally efficient approach for unsupervised mobile user behavior anomaly detection within the scope of this study.

The increasing complexity and scale of modern network traffic driven by IoT and cloud-based infrastructures have made accurate intrusion detection a critical challenge. Conventional network intrusion detection systems (NIDS) and many deep learning–based approaches struggle to reliably detect minority and stealthy attacks due to severe class imbalance and limited discrimination of subtle traffic patterns. To address these limitations, this study proposes a hybrid CNN–RBF–Attention framework for network intrusion detection. The proposed model integrates three complementary components: (i) a convolutional neural network for hierarchical feature extraction from network flow data, (ii) a radial basis function (RBF) network for localized nonlinear classification using prototype-based decision regions, and (iii) an attention mechanism that adaptively weights RBF activations to emphasize discriminative traffic patterns. SMOTE is applied exclusively to the training data to mitigate class imbalance. The framework is evaluated on the widely used CICIDS2017 and CICIDS2018 benchmark datasets in both binary and multiclass settings, using recall, precision, F1-score, confusion matrices, and ROC analysis. Experimental results demonstrate that the proposed hybrid model consistently outperforms standalone CNN and RBF baselines, particularly in terms of recall and F1-score. On the CICIDS2018 dataset, the model achieves 99.81% accuracy and 99.81% F1-score in binary classification, and 99.54% accuracy and 99.54% F1-score in multiclass classification. On CICIDS2017, it achieves 98.12% accuracy and 98.12% F1-score in binary classification, and 98.92% accuracy and 98.92% F1-score in multiclass classification. Confusion matrix and ROC analyses further show strong class separability and reliable performance in low–false-positive-rate regions, which is critical for real-world IDS deployment. These results confirm that combining deep hierarchical feature learning, localized prototype-based classification, and attention-guided refinement yields a robust, operationally reliable intrusion detection framework for highly imbalanced network environments.

This study presents a comprehensive comparative analysis of four traditional machine learning algorithms Decision Tree, Random Forest, K-Nearest Neighbors, and Support Vector Machine for Android malware detection using the preprocessed TUANDROMD dataset comprising 4,465 instances and 241 features representing both static and dynamic application characteristics. Motivated by the limitations of conventional signature-based and hybrid detection methods, especially in managing imbalanced datasets and detecting emerging malware variants, the study employed SMOTE to ensure balanced training data and fair model evaluation. The dataset was divided into 80% training and 20% testing subsets, and models were assessed using key performance metrics including accuracy, precision, recall, F1-score, and ROC AUC. The findings revealed that the proposed Random Forest model outperformed the other classifiers, achieving an accuracy of 0.993, precision of 0.992, recall of 1.000, F1-score of 0.996, and a near-perfect ROC AUC of 0.9998 surpassing state-of-the-art approaches. These results affirm the superior predictive capability, consistency, and robustness of the Random Forest algorithm in Android malware detection. The study concludes that base models, when integrated with class-balancing techniques, provide reliable and efficient malware detection across imbalanced datasets. For future research, the study recommends exploring advanced hybrid or ensemble frameworks that integrate Random Forest with deep learning architectures or other meta-heuristic optimization techniques to further enhance detection accuracy, adaptability, and resilience against rapidly evolving Android malware threats.

The integration of Artificial Intelligence (AI) into precision agriculture has significantly improved plant disease recognition; however, many existing deep learning models remain computationally expensive and feature-redundant, limiting their deployment on low-power and edge devices. To address these limitations, this study proposes a lightweight framework for maize leaf disease recognition based on serial deep feature extraction, dimensionality reduction, and machine-learning–based classification. A pre-trained MobileNetV2 network is employed as a fixed feature extractor to obtain discriminative visual representations, while Principal Component Analysis (PCA) is applied to reduce feature dimensionality by approximately 76%, retaining 95% of the original variance and improving computational efficiency. The compressed features are subsequently classified using a Radial Basis Function Support Vector Machine (RBF-SVM), optimized via grid search and cross-validation. Experiments conducted on a four-class maize leaf disease dataset (Northern Leaf Blight, Common Rust, Gray Leaf Spot, and Healthy), with class imbalance handled during training, demonstrate that the proposed MobileNetV2–PCA–SVM pipeline achieves 97.58% accuracy, 96.60% precision, 96.59% recall, and 96.59% F1-score, outperforming the DenseNet201 + Bayesian-optimized SVM baseline (94.60%, 94.40%, 94.40%, and 94.40%, respectively). This improvement corresponds to a 2.98% accuracy gain, a 55% reduction in error rate, an 86% reduction in model parameters (20.31M to 2.75M), and an 85% reduction in model size (81 MB to 12 MB). These results indicate that the proposed framework provides a compact and efficient solution with strong potential for deployment in resource-constrained agricultural environments.

Students with deaf disabilities often have difficulty understanding verbal instruction, but have significant advantages in visual and motor aspects. This service activity aims to apply multisensory methods through Suminagashi craft art training (cloth marbling) to improve students' fine motor skills and creativity at SLB B Yakut Purwokerto. The implementation method uses a descriptive qualitative approach with a "see-learn-do" strategy involving 20 high school students. The learning process integrates visual stimulation through structured demonstrations and tactile stimulation through the exploration of material textures. The results of the activity showed that the multisensory approach was effective in bridging the barriers of deaf students with disabilities in communicating due to hearing limitations. Students demonstrate a high level of visual focus and are able to replicate marble motif making techniques with precision without relying on complex verbal explanations. The combination of cue instructions and direct touch experiences has been proven to minimize miscommunication and improve understanding of abstract concepts of the material. This activity recommends the use of visual-tactile strategies as an adaptive inclusive learning method to support the vocational independence of deaf people.          

The maritime education sector faces complex challenges in preparing competent seafarers amid the rapid advancement of digital technology. This study investigates the effect of artificial intelligence-based simulations and AI-based competency assessments on competency achievement levels among nautical cadets at Indonesian maritime training institutions. The research design employed a convergent parallel mixed-methods approach, integrating quantitative and qualitative methods to gain a comprehensive understanding. Quantitative data were collected from 150 cadets using a validated questionnaire. In comparison, qualitative data were obtained through in-depth semi-structured interviews with fifteen instructors and ten cadets. Multiple regression analysis revealed that the research model significantly predicted cadet competency achievement. The findings indicate that AI-based assessments exert a stronger influence than AI simulations in improving competency. The qualitative exploration highlighted adaptive feedback mechanisms and personalized learning pathways as critical success factors in implementing learning technologies. This study provides empirical evidence for maritime institutions to prioritize strategic investments in AI-based assessment systems while maintaining a human-centered pedagogy. The research contribution lies in integrating fourth industrial revolution technologies into the training, certification, and watchkeeping standards compliance framework for seafarers, thereby strengthening Indonesia's maritime education ecosystem and aligning it with international standards.

Hypertension is a major non-communicable disease, particularly challenging in regions with extensive service areas. Community health volunteers are essential for prevention and management through blood pressure measurement. Existing training programs focus primarily on knowledge, highlighting the need to integrate cognitive learning with small-group skills practice to enhance practical competencies and community-based hypertension control. A quasi-experimental design with a pretest–posttest design was conducted to evaluate the effectiveness of combined lecture and small-group training. Knowledge was assessed before and after training, while skills were evaluated post-intervention. Thirty volunteers from the Public Health Center Selakau participated. The results showed that knowledge of blood pressure measurement improved significantly, with pretest scores of 74.67 ± 16.34 rising to posttest scores of 90.00 ± 10.50 (p < 0.005). Posttest evaluation of practical skills showed a mean score of 80.93 ± 13.35, indicating proficient performance in most assessed items. Combined lecture and small-group training effectively enhanced both knowledge and practical skills of community health volunteers in blood pressure measurement. Integrating cognitive learning with hands-on practice strengthens theoretical understanding and field competencies, supporting more effective community-based hypertension control programs.

Children with special needs require specialized educational services that accommodate their diverse learning characteristics, and digital-based media offers flexible and adaptive learning opportunities tailored to individual abilities. The digital learning media “Voca Poly” was developed to provide education and practical skills training for blind and physically disabled children, particularly in strengthening knowledge and cooking skills to support independent living. This community service program was implemented through three main stages: preparation, implementation, and evaluation. The preparation stage involved needs assessment, media design, and coordination with partner institutions. The implementation stage included interactive learning sessions using the Voca Poly educational game, focusing on food security concepts through English vocabulary from planting to food processing. The evaluation stage assessed participants’ understanding, engagement, and skill improvement. The results showed increased knowledge of cooking concepts and improved practical skills among participants. Moreover, the game-based approach successfully fostered enthusiasm, motivation, and active participation among visually impaired and physically disabled children.

This study aims to examine the role of mathematics education in building numerical literacy in elementary schools, as well as strategies that can be implemented to enhance students' numerical skills. Numerical literacy is a fundamental skill needed by students to understand, interpret, and use numerical information in daily life. The research method used is library research, reviewing various related literature, including books and recent scholarly journals in the field of mathematics education. The results show that the integration of technology, the use of contextual approaches, and collaborative learning are effective strategies in improving students' numerical literacy. Technology helps simplify the understanding of abstract concepts, while contextual and collaborative approaches make mathematics more relevant and applicable. The implications of these findings are the need for further development in teacher training to integrate technology in teaching and the importance of involving students in more interactive and contextual learning processes. This study also suggests further evaluation of the implementation of these strategies in various educational settings.

This study aims to compare the performance of Multiple Linear Regression (MLR) and Random Forest Regression (RFR) in predicting student performance based on academic scores. Student performance is defined as the average of math scores, Reading Scores, and writing scores. This study uses a quantitative approach with a comparative design based on predictive modeling. The data used is secondary data from the Student Prediction dataset obtained through the Kaggle platform, which was processed using the Python programming language through the Google Colab platform. The analysis stages included the formation of performance variables, the separation of training and test data with a ratio of 80:20, model training, and evaluation using the Mean Squared Error (MSE), Mean Absolute Error (MAE), and coefficient of determination (R²) metrics. The results show that the Multiple Linear Regression model produced an MSE value of 2.74 × 10⁻²⁸, an MAE of 1.51 × 10⁻¹⁴, and an R² of 1.000. Meanwhile, Random Forest Regression produced an MSE of 0.296, an MAE of 0.375, and an R² of 0.998. These findings indicate that both models have a very high level of accuracy, but Multiple Linear Regression provides the best performance. This is due to the strong linear relationship between the input variables and the target variables formed directly from the combination of academic values. Thus, the linear regression model is proven to be more suitable for use in data structures that have simple linear relationships compared to ensemble-based models.

This study aims to strengthen students’ networking competencies through the implementation of straight and cross LAN cable assembly training at SMKN 53 Jakarta. The background of this research is based on the need to enhance students’ practical skills in basic computer networking, particularly in understanding cable configurations and applying crimping techniques according to industry standards. Many students experience difficulties in differentiating wiring standards and applying correct crimping procedures, which impacts their readiness for industry practice. This research employed a practical training approach combined with demonstration and hands-on methods. The participants were students of the Computer and Network Engineering program. Data were collected through observation, performance assessment, and competency tests before and after the implementation. The findings indicate a significant improvement in students’ understanding of cable color standards (T568A and T568B), accuracy in assembling straight and cross cables, and testing results using LAN testers. Students demonstrated higher levels of technical accuracy, problem-solving skills, and work discipline after the intervention. The implementation of structured practical activities proved effective as a medium for strengthening networking competencies. The study implies that continuous practice-based learning aligned with industry standards is essential in vocational education to improve students’ technical readiness and employability in the networking field.

The Public Speaking and Computer Network Cabling Training Activity at SMP YPK Sele Be Solu aims to improve students' competence in communication and technical skills in computer networking. The activity was carried out using an experiential learning approach with stages of observation, socialisation, hands-on training, and evaluation. The results of the activities show an increase in students' confidence in public speaking and basic technical skills in network cabling, including LAN cable installation, RJ-45 connectors, and cable testing. Intensive mentoring during the training helped students overcome difficulties and optimise their understanding. This activity demonstrates that combining soft skills and hard skills training through direct experience can improve students' overall competence, motivate learning, and prepare students to face academic and digital challenges in the future. In addition, this activity also strengthened cooperation between schools and universities in supporting the improvement of skills-based education quality. The enthusiasm of students during the training showed that practice-based learning methods are very effective at the junior high school level. With this activity, students have a strong foundation to develop their interests and potential in the fields of communication, information technology, and readiness to face the rapidly developing digital world.

This community service activity aims to strengthen students’ soft skills and hard skills through public speaking and computer network cabling training at SMA Negeri 1 Sorong City. The activity's background is rooted in the state of computer learning, which still focuses on the introduction of software and hardware, while communication skills and computer network practices have not been fully developed. The method for implementing the activity uses an experiential learning approach, comprising the planning, socialization, training, and evaluation stages. Soft skills training is focused on improving public speaking skills through public speaking practice, while hard skills training is carried out through direct practice of computer network cabling using UTP cables and supporting devices. The results of the activity show an increase in students’ communication skills and self-confidence, as well as an increase in students' understanding and technical skills in computer network cabling. This activity provides an applicable learning experience and strengthens the synergy between universities and high schools in developing student

Customer churn is a strategic challenge for digital streaming platforms because it directly Impacts revenue and business sustainability. This study aims to analyze the factors influencing customer Churn and develop a churn prediction model using the Random Forest algorithm. The study uses a Quantitative approach with an explanatory design and utilizes secondary data from the Netflix Customer Churn and Engagement Dataset available on Kaggle. The dataset consists of 1,000 customer data with 16 Variables covering demographic characteristics, service usage behavior, financial condition, and customer Satisfaction level. The data was processed through preprocessing, one-hot encoding, and a 70:30 split Between training and test data. Model performance was evaluated using accuracy, precision, recall, F1 Score, and ROC-AUC metrics. The results show that the Random Forest model produces an accuracy of 53.7%, precision of 56.3%, recall of 63.6%, F1-score of 59.7%, and ROC-AUC of 0.534, indicating Moderate predictive ability and only slightly better than random classification. Feature importanceAn.evealed that user engagement levels, such as viewing duration and frequency of interactions, Were the most dominant factors influencing churn, followed by economic factors and customer satisfaction. The results of this study are expected to provide a basis for streaming platforms to design more effective Customer retention strategies.

A student's learning success is largely determined by their academic evaluation. Estimating a student's final grade can assist educational institutions in conducting initial assessments of academic achievement. This study aims to analyze the performance of the Multiple Linear Regression (MLR) and Random Forest (RF) algorithms in predicting students' final grades using Google Colab. This research method uses a quantitative approach using secondary data that includes age, mid-term exam scores, final exam scores, and categorical variables as independent variables, with the final grade as the dependent variable. The research process is carried out through data preprocessing steps, dividing training and test data, model training, and performance evaluation using Mean Squared Error (MSE), Mean Absolute Error (MAE), and the coefficient of determination (R2). The results show that the Random Forest algorithm provides more accurate prediction accuracy compared to the Multiple Linear Regression algorithm, especially in identifying nonlinear relationships between variables. Therefore, the Random Forest algorithm is more recommended for predicting students' final grades with complex data characteristics.

International student enrollment is a critical driver of financial sustainability for Higher Education Institutions (HEIs). While advanced forecasting is standard in the corporate sector, its application in educational planning remains limited. This study addresses this gap by comparing the predictive performance of ARIMA, Facebook Prophet, and Long Short-Term Memory (LSTM) models. Using a publicly available annual dataset from a US-based institution (2000–2022), the analysis employed a strategic partition training on 2000–2017 and testing on 2018–2019 to validate models on stable, pre-pandemic data. Empirical results revealed that the statistical ARIMA (2,1,0) model demonstrated superior accuracy, achieving a Mean Absolute Percentage Error (MAPE) of 1.26%. Conversely, Prophet (11.81%) and LSTM (13.84%) struggled with the limited sample size, failing to generalize effectively compared to the linear approach. The findings suggest that for annual enrollment trends, parsimonious statistical models outperform complex deep learning architectures, providing administrators with a robust, accessible framework for data-driven strategic decision-making.

This training aims to equip students of to enhance the graphic design skills of students at SMK XI DKV Pangeran Antasari through Adobe Photoshop training. This activity is conducted in response to the growing need for graphic design proficiency among students in the digital age. This service adopts a practical qualitative research method with a direct approach, enabling participants to immediately apply the acquired knowledge. The methods employed include interactive workshops, practical demonstrations, and skills evaluation. The results of the activity demonstrate a significant improvement in students' abilities, with 85% of participants capable of producing simple designs independently. It is therefore expected that upon completion of this training, participants will be able to enhance their creativity in learning, particularly in the field of graphic design, as a primary foundation for becoming professional designers in the workforce."

This study aims to describe the implementation of the Al-Miftah method in improving the mastery of reading traditional Islamic texts at Al-Mashhad Senior High School in Cijurai, Sukabumi. This study uses a qualitative approach with a descriptive research type. Data collection techniques were carried out through observation, in-depth interviews, and documentation of Al-Miftah teachers, students, and related parties. The results of the study indicate that the Al-Miftah method is implemented through the stages of introducing the basic rules of nahwu-sharaf, practicing reading traditional Islamic texts in stages, and strengthening understanding through repetition and direct practice. The implementation of this method has been proven to help students understand the structure of Arabic and improve their ability to read traditional Islamic texts, although several obstacles were still encountered, such as differences in students' Arabic language backgrounds, limited learning time, minimal independent practice, and limited mastery of the method by teachers. Efforts made to overcome these obstacles include learning grouping, time optimization, out-of-class practice assistance, and improving teacher competency through continuous training and evaluation. Thus, the Al-Miftah method has a positive contribution in increasing mastery of reading traditional books at Al-Mashhad Cijurai Sukabumi High School if supported by appropriate and continuous learning strategies

The development of education in the digital era requires students to master not only academic skills, but also soft skills and hard skills as a provision to face the challenges of education and the world of work. However, the learning process at Papua High School in Sorong City is still teacher-oriented, so students' communication courage and technological skills have not developed optimally. This community service activity aims to empower students through public communication training (public speaking) and computer network cable techniques to strengthen 21st century skills. The method used is a participatory approach with hands-on practice through four stages: observation, socialization, training, and evaluation. Public speaking training is focused on increasing students' courage and confidence in communicating, while network cable training emphasizes basic technical skills such as recognizing cable types, RJ-45 connector functions, crimping techniques, and connectivity testing. The results of the activities showed high participation and enthusiasm from students, as well as increased public speaking courage and basic understanding in network cable making. The integration of public communication training and technology skills has been proven to have a positive impact on the development of student competencies. Therefore, this activity is recommended as a model for developing soft skills and hard skills for high school students, especially in areas with limited access to technology learning.

Emotional intelligence (EI) is an essential competency for early childhood education (ECE) teachers in fostering supportive and emotionally responsive learning environments. However, many ECE teachers in Indonesia have limited access to structured EI training. This community service program aimed to strengthen the emotional intelligence of non-formal ECE teachers from HIMPAUDI Cipayung, East Jakarta, through interactive lectures and reflective practice using Points of You (Flow) cards. The training focused on four EI domains: self-awareness, self-management, social awareness, and relationship management. Results showed a substantial improvement in participants’ understanding, with average scores increasing from 30.5 (pre-test) to 90.9 (post-test). Most participants (85%) reported that the training was beneficial and expressed readiness to apply the strategies in classroom practice. Observations also indicated improved teacher confidence in managing children’s emotions and building positive interactions. These findings demonstrate that reflective and experiential EI training effectively enhances ECE teachers’ emotional competencies and supports the creation of a positive social-emotional learning environment.