Publication Search

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.

Prompt injection has emerged as a critical security threat for Large Language Models (LLMs), exploiting their inability to separate instructions from data within application contexts reliably. This paper provides a structured review of current attack vectors, including direct and indirect prompt injection, and highlights the limitations of existing defenses, with particular attention to the fragility of Known-Answer Detection (KAD) against adaptive attacks such as DataFlip. To address these gaps, we propose a novel, hybrid, multi-layered detection framework that operates in real-time. The architecture integrates heuristic pre-filtering for rapid elimination of obvious threats, semantic analysis using fine-tuned transformer embeddings for detecting obfuscated prompts, and behavioral pattern recognition to capture subtle manipulations that evade earlier layers. Our hybrid model achieved an accuracy of 0.974, precision of 1.000, recall of 0.950, and an F1 score of 0.974, indicating strong and balanced detection performance. Unlike prior siloed defenses, the framework proposes coverage across input, semantic, and behavioral dimensions. This layered approach offers a resilient and practical defense, advancing the state of security for LLM-integrated applications.

Improper water management in rice cultivation can lead to water stress, which reduces productivity. Conventional monitoring has limitations on large-scale lands, necessitating more efficient remote sensing technologies. This study aims to develop a water stress identification system for rice plants in the late vegetative phase using multispectral drone imagery integrated with an Artificial neural network (ANN). The research method employs an experimental approach with six water availability levels in Karyamukti Village, Sumedang. Field reference data were obtained through soil moisture sensors converted into Available Water (AW) values. Image processing stages included orthomosaic reconstruction, leaf object segmentation, and transformation of vegetation indices (NDVI, NDRE, GNDVI, etc.) as model inputs. The results show that the ANN model with a four-hidden-layer architecture achieved training and validation accuracies of 94–95%. In the independent testing phase, the model produced an accuracy of 94.60% with an F1-Score of 93.33%. Spatial visualization of the prediction results indicates a consistent water condition distribution across rice plots. In conclusion, the integration of multispectral drones and ANN provides an accurate non-destructive solution for spatial monitoring of water availability in rice plants.

Private vehicles are a frequently used mode of transportation because they are considered more practical. However, using private vehicles carries several risks, such as traffic accidents due to drivers losing focus on the road due to other activities, such as making calls on smartphones, drinking, or operating the radio. Approximately 90% of accidents are caused by human error. Convolutional Neural Network (CNN) is a type of neural network commonly used on image data. CNN is often used for image classification due to its high performance and accuracy. Therefore, this study aims to analyze the performance of CNN for the classification of distracted driving activities. The results show that the CNN model is able to effectively classify images of distracted driving activities, with an accuracy of approximately 99% across all datasets and across all input image size variations. Furthermore, the results of this study also show that differences in right-hand and left-hand drive datasets do not significantly affect model accuracy. Variations in input image size also do not significantly affect model accuracy, but do affect the training duration.

This study aims to analyze the strategies of tahfidz teachers in improving students’ Qur’anic memorization quality through a literature review approach. Memorization quality is not merely measured by the quantity of verses memorized, but also includes accuracy of recitation, fluency, consistency in muroja’ah (revision), and long-term memory retention. This research employed a literature review method by examining national and international journal articles as well as relevant academic books published between 2020 and 2025. Data were analyzed using content analysis techniques to identify themes, patterns, and research gaps related to teachers’ strategies in tahfidz learning. The findings reveal that effective tahfidz teaching strategies can be classified into four main aspects: (1) structured and consistent implementation of repetition (tikrar), (2) reinforcement of muroja’ah and periodic evaluation, (3) motivational strategies and character development, and (4) innovative learning approaches integrating collaboration and educational technology. The tikrar method is proven effective in strengthening memorization retention when supported by systematic program planning. Furthermore, intrinsic motivation, a conducive learning environment, and varied instructional methods significantly contribute to maintaining students’ memorization stability. In conclusion, improving the quality of Qur’anic memorization depends not only on repetition frequency but also on the integration of pedagogical strategies, affective-spiritual approaches, and instructional innovation. This study provides a conceptual contribution to the development of more comprehensive and sustainable tahfidz learning strategies.

Indramayu mango is a seasonal fruit that is highly favored due to its delicious taste and high nutritional content. However, high mango production is often not supported by adequate post-harvest facilities, particularly in terms of fruit ripeness classification. Currently, mango ripeness classification is still performed manually, which tends to be subjective and inconsistent. To address this issue, this study proposes a ripeness detection system for Indramayu mangoes by integrating the TGS2602 gas sensor and the YOLOv11 algorithm based on image processing. The TGS2602 sensor is used to detect ethylene gas emitted by ripe mangoes, while YOLOv11 is employed for visual image analysis of the fruit. This study aims to evaluate the system’s performance in classifying ripe and unripe mangoes, as well as analyze the integration between the gas sensor and the object detection model. The test results show that the TGS2602 sensor can detect increased ethylene gas concentration in ripe mangoes, while YOLOv11 demonstrates high accuracy in detecting mangoes based on visual images, with precision and recall close to 1.0. The system was also tested under various lighting conditions, including dark environments, and still performed well, although with a slight decrease in accuracy under low-light conditions.

Freshwater fish diseases significantly affect aquaculture productivity and economic sustainability, while accurate visual classification remains challenging due to interclass similarity and image variability. This study presents a comparative evaluation of three deep learning architectures—DenseNet201, ResNet50, and EfficientNetV2-S—using a stepwise optimization strategy combined with Gradient-weighted Class Activation Mapping (Grad-CAM) for freshwater fish disease classification. Models were trained through three phases: baseline, optimized, and fine-tuned. Performance was evaluated using accuracy, precision, recall, F1 score, Matthews correlation coefficient (MCC), Cohen’s kappa, and per-class ROC–AUC. Results show consistent performance improvement across all architectures, with EfficientNetV2-S achieving the highest accuracy (97.14%), followed by ResNet50 (96.11%) and DenseNet201 (94.40%). High ROC–AUC values (>0.98) indicate strong discriminative capability. Grad-CAM analysis confirms that all optimized models focus on biologically relevant lesion regions, enhancing model transparency and reliability.

The rapid growth of e-commerce in Indonesia, particularly the Shopee platform, has generated a large volume of user reviews on the Google Play Store, which can be analyzed to understand consumer sentiment. This study aims to compare the performance of the Support Vector Machine (SVM) and Random Forest (RF) algorithms in binary sentiment classification (positive and negative) on Shopee reviews, as well as to statistically test the significance of their differences using One-Way ANOVA. A total of 400,498 reviews were collected via web scraping, preprocessed through text normalization, tokenization, and Indonesian language stemming, and then feature-extracted using TF-IDF and Count Vectorizer. Evaluation results show that SVM achieved an accuracy of 91.77%, precision of 91.49%, recall of 91.77%, and F1-Score of 91.56%, while RF achieved an accuracy of 90.07%, precision of 91.68%, recall of 90.07%, and F1-Score of 90.55%. ANOVA confirmed that the performance difference between the two algorithms is statistically significant (p-value = 0.0007) with a large effect size (η² = 0.1815). Therefore, SVM is recommended as a more optimal and consistent algorithm for automated sentiment analysis of Indonesian e-commerce reviews, while also providing a replicable methodological framework for similar future research.

The spread of hoax news in digital media is a serious problem because it can affect public opinion and social stability. This study aims to classify hoax news using the Support Vector Machine (SVM) algorithm. The dataset used is a hoax clarification dataset from the Ministry of Communication and Digital (Komdigi) of the Republic of Indonesia, totaling 1,872 data. The research process includes data collection, text pre-processing, feature extraction using TF-IDF, and classification using the SVM algorithm. Implementation was carried out using Google Colaboratory (Google Colab). Test results show that the SVM algorithm is able to provide good performance in classifying hoax news based on its topic with satisfactory accuracy, precision, recall, and F1-score values.

This study evaluates public sentiment toward Constitutional Court Decision No. 90/PUU-XXI/2023 regarding the age limit for presidential and vice-presidential candidates, a controversial issue closely related to Indonesia’s democratic dynamics. Understanding public opinion on Twitter, as a major platform for political expression, is essential for informing electoral policy formulation. Data were collected using Tweet Harvest through Google Colab and analyzed using the Naïve Bayes algorithm as the primary sentiment classification method, with RapidMiner employed to support and streamline the analytical process. The analysis process included data cleaning, text normalization, stopword removal, manual labeling of 80 tweets as training data, and automatic sentiment classification to identify positive and negative sentiments. From a total of 151 analyzed tweets, 84 (55.63%) were classified as negative and 67 (44.37%) as positive, with the model achieving an accuracy of 66.67%. These findings suggest a tendency toward public opposition to the decision, reflecting dissatisfaction among Twitter users. The study demonstrates that Naïve Bayes is reasonably effective for sentiment classification with limited datasets and provides insights for policymakers in understanding public responses to election-related regulations.

This research develops a machine learning model to classify customer loyalty using the Random Forest algorithm. Customer churn is a critical issue that reduces revenue and increases acquisition costs. A dataset of 50,000 customers from global e-commerce and subscription platforms was processed through data cleaning, imputation, outlier handling, and class balancing with SMOTE. The Random Forest model was built as a baseline and optimized with hyperparameter tuning. Evaluation using accuracy, precision, recall, and F1-score shows that the optimized model achieved 90.81% accuracy and 83.87% F1-score, outperforming previous Naïve Bayes approaches. Feature importance analysis highlights customer service interactions, lifetime value, and demographic factors as key predictors of churn. These findings demonstrate Random Forest’s effectiveness in churn prediction and provide practical insights for customer retention strategies

Since the announcement of the policy plan for a name transfer system in the sale of used mobile phones, the issue has attracted widespread public attention and discussion. People have expressed their opinions on social media platforms, particularly TikTok. This study aims to classify the sentiment of TikTok users using Naive Bayes and Support Vector Machine (SVM) algorithms. The data were collected through a comment scraping technique on related content.The research stages include text preprocessing, sentiment labeling into positive, negative, and neutral categories, and feature extraction using TF-IDF. The classification process employs Naive Bayes and Support Vector Machine algorithms, which are then evaluated based on accuracy, precision, recall, and F1-score. The results of this study indicate that both methods are capable of classifying sentiment effectively. However, the Support Vector Machine method is superior to the Naive Bayes method with an accuracy rate of 99.57% compared to 94.30%. This study is expected to help the government understand public responses to the planned policy of the used mobile phone name transfer system.

The management of kWh meter replacement data at PLN ULP Ampera Palembang is still largely handled through manual recording, which often causes data inconsistencies and delays in monitoring activities. This study aims to design an information system that supports the monitoring of kWh meter replacement data based on actual user needs. The research applies a descriptive qualitative method using the User Centered Design (UCD) approach, where users are actively involved throughout the design process. The stages include understanding the work context, identifying user requirements, developing system design solutions, and evaluating the proposed design. The outcome of this research is a kWh meter data monitoring system design that is expected to facilitate data management, improve accuracy, and support more efficient monitoring processes.

Open-pit mining activities in Indonesia contribute significantly to the national economy but require stringent monitoring to mitigate environmental degradation. Conventional monitoring methods relying on terrestrial surveys are often constrained by vast coverage areas, high operational costs, and limited field accessibility. This study aims to develop an artificial intelligence model capable of automatically detecting and mapping mining areas to enhance surveillance efficiency. The applied method is Deep Semantic Segmentation utilizing the U-Net Convolutional Neural Network (CNN) architecture. The model was trained using Sentinel-2 satellite imagery, focusing exclusively on Red, Green, and Blue (RGB) spectral channels to replicate human visual perception. Experimental results demonstrate that the proposed model performs reliable segmentation of mining areas, achieving an Accuracy of 93.58% and a Global Intersection over Union (IoU) of 0.8067. These findings indicate that the U-Net architecture can effectively extract spatial features of mines even when utilizing standard visual data. This research contributes to the development of an efficient, cost-effective, and scalable digital monitoring prototype to support innovation in sustainable environmental governance.

This study discusses clause structures in online news articles published by Detik.com, highlighting the issue of UNNES student Iko Juliant Junior in September 2025. The background of this research lies in the importance of understanding how digital media utilize language structures to convey information effectively, concisely, and objectively. In the context of online journalism, the selection and arrangement of clauses play a crucial role in shaping clarity, focus, and meaning as perceived by readers. The purpose of this study is to describe the forms and types of clauses used in news texts and to explain their functions in maintaining the accuracy and credibility of information delivery. The research employs a qualitative descriptive method with a focus on syntactic studies. The data consist of clauses found in Detik.com news articles published in September 2025, analyzed based on their syntactic components: subject, predicate, object, complement, and adverbial. The findings show that the clause patterns of Subject–Predicate (S–P) and Subject–Predicate–Object (S–P–O) are the most dominant, with verbal predicates appearing most frequently. These patterns reflect the efficient, factual, and easily understood style of online journalistic language. This study contributes to the development of syntactic research and serves as a practical reference for journalists and readers in understanding effective language use in digital media.

Language is a means of communication between individuals and groups to convey ideas, concepts, or insights. Language can be expressed orally or in writing. Spoken language is temporary or contemporary, while written language is permanent. Spoken language is delivered without an intermediary media, whereas writing serves as an academic archive to ensure that the conveyed insights are preserved. Therefore, writing is done, for example, in encyclopedias, which contain collections of information about science, knowledge, history, arts, technology, and others in a broad and easily understandable manner. Writing books is very important in delivering accurate and easily understood information, so the use of language must adhere to linguistic rules. However, errors are often found that can disrupt clarity and accuracy of information. This research aims to identify types of sentence errors in the encyclopedia book titled “Islamic Kingdoms in Indonesia.” The study examines the use of language in accordance with linguistic rules using a qualitative method and a descriptive approach, analyzing data from various chapters in the book in depth. The results show errors in sentence structure, punctuation, and incorrect terminology, which can reduce the quality of the information. The benefit of this research is to provide recommendations for writers and publishers to improve the encyclopedia so that it becomes more informative and easier to understand, as well as to enhance the quality of scholarly works in the field of historiography.

This study aims to examine the role of Computer-Assisted Audit Techniques (CAATs) in evaluating internal control within accounting information systems (AIS) to detect fraud in the expenditure cycle. The research employs a literature review method by analyzing five relevant studies selected based on publication criteria within the last ten years and a focus on technology-based auditing, internal control, and fraud. The findings indicate that CAATs, through features such as test data and parallel simulation, are effective in identifying system weaknesses, detecting transaction anomalies, and strengthening controls in the expenditure cycle. Fraud in this cycle is commonly caused by weak authorization, incomplete documentation, and expenditures conducted without proper procedures. CAATs address these challenges through data-driven and automated audit approaches. In conclusion, CAATs represent a strategic solution for enhancing monitoring accuracy, preventing fraud, and supporting organizational transparency and accountability in the digital era.

This study aims to describe the structure, syntactic function, and semantic role of the nominal clause in JawaPos.com news texts published in August 2025. By using a descriptive qualitative approach, this study applied syntactic analysis to explore how the nominal clause operates within the linguistic and contextual framework of media discourse. Data were collected from online news articles and analyzed using Indonesian syntactic and semantic theory. The results show that nominal clauses play a crucial role in constructing discourse structure and meaning through their syntactic functions as subjects, objects, complements, and adverbs. In semantics, nominal clauses perform various roles, including as agents, objects, recipients, and markers of condition or location. The pattern of nominal clause use reflects the concise, efficient, and communicative nature of journalistic language. The correlation between syntactic function and semantic role shows a cohesive relationship that enhances clarity and accuracy in conveying information. This study contributes to the development of syntactic and media linguistic research and serves as a valuable reference for journalists and language educators in promoting effective and ethical language use in journalism.

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 analyze the performance of Android-based attendance applications used at PT PLN ULP West Sumba by integrating two evaluation methods, namely WebQual 4.0 and Importance Performance Analysis (IPA). This attendance application functions to monitor employee attendance digitally so that it is expected to be able to improve the efficiency and accuracy of data recording. Evaluations are conducted to assess the extent to which the application meets the needs of users as well as the expected performance. WebQual 4.0 is used to measure the quality of user experience in terms of ease of access, interactivity, trust, and satisfaction, while IPA is used to compare the level of user interest with application performance based on four main attributes: system quality, information quality, service quality, and usage quality. The research data was collected through a survey with questionnaires compiled according to WebQual 4.0 and IPA indicators, involving application users at PT PLN ULP West Sumba. The results show that the majority of users are satisfied with the ease of use and performance of the application, but there are aspects that need to be improved, especially the speed of the system and a more user-friendly interface design. The science analysis emphasizes that the quality of systems and information is a crucial factor that must be a priority for development. This research provides strategic recommendations for PT PLN ULP West Sumba to improve the performance of the attendance application and support the company's operational needs in a sustainable manner.