Publication Search

This study applies the Mean Variance model, which aims to form an optimal portfolio composition in the health, property, and cyclical consumer sectors and combine the three sectors into one portfolio, then visualize its efficient frontier. This study analyzes the return profiles and compares the risks of each portfolio using alternative risk measures such as the Coefficient of Variation (CV), Value at Risk (VaR), and Conditional Value at Risk (CVaR). Daily closing price data for the three sectors listed on the Indonesia Stock Exchange (IDX) from March 2, 2020, to March 3, 2025, were used in this study. Stock selection was conducted using purposive sampling, followed by selecting seven stocks for optimization based on the lowest Coefficient of Variation (CV) value. Portfolio optimization analysis was conducted using the Python programming language with Visual Studio Code software. The findings of this study indicate that the combined portfolio incorporating the three sectors is the most efficient, with an expected return of 0.104%, standard deviation of 0.007, and alternative risk measures such as Coefficient of Variation (CV) 6.9328, Value at Risk (VaR) of -0.99%, and Conditional Value at Risk (CVaR) of -1.44%, which are lower than those of single-sector portfolios. Visualization of the efficient frontier curve confirms that the combined portfolio offers better results in terms of risk and return. The results of this study indicate that cross-sector diversification can significantly reduce risk and prevent significant losses.

The use of protective equipment in the form of helmets is an important aspect of ensuring motorcycle rider safety. However, violations of helmet usage still frequently occur and are difficult to monitor continuously. This study proposes a real-time helmet detection system using the YOLOv8 object detection method. The YOLOv8n model was trained using a helmet and no-helmet image dataset that underwent data augmentation to improve the model’s robustness against variations in environmental conditions. The system was implemented using the Python programming language with the support of the Ultralytics and OpenCV libraries. The system input was obtained from a webcam with a resolution of 640×640 pixels, where each video frame was processed in real time to detect the Helmet and No Helmet classes. The system displays bounding boxes and class labels in real time and is equipped with a violation duration calculation mechanism. When a no-helmet condition is detected continuously, the system generates pop-up alerts and automatic notifications via the Telegram application. The experimental results show that the system is capable of detecting helmet usage and no-helmet violations in real time with stable performance. The integration of violation duration calculation helps reduce momentary detection errors and improves the reliability of identifying valid violations

Diabetes Mellitus is a group of metabolic diseases characterized by hyperglycemia resulting from defects in insulin secretion, insulin action or both. Hyperglycemia is a medical condition in the form of an increase in glucose levels beyond normal limits which is a characteristic of several diseases, especially Diabetes Mellitus, in addition to various other conditions. Diabetes Mellitus is currently a global health threat. Classification is one of the techniques of data mining that can be used to help predict the results of the classification of types of diabetes using the naïve Bayes algorithm. Testing was carried out using 5 evaluation models including rapid miner with 3 options, namely use training set, 5 Fold Cross-Validation, 10 Fold Cross-Validation, and 2 other evaluation models, namely Microsoft Excel and Python. Testing data regarding Diabetes Mellitus has high accuracy in the excel evaluation model, which is 89.00% compared to other evaluation models. Meanwhile, the lowest accuracy is the Python evaluation model which obtains an accuracy of 86.36%. The Naïve Bayes algorithm can be said to be one of the most effective algorithms, both in terms of calculations and the final results, where the test can be used as a basis for diabetes mellitus considering the accuracy results are above 85%.

Diabetes Mellitus is a disease caused by the failure of the pancreas organ in producing the hormone insulin in excess causing increased blood sugar levels and resulting in a lack of insulin. This study discusses the application of the k-means clustering method to determine risk factors for diabetes mellitus. By using the clustering method, data will be grouped into several clusters or groups which in this study compare by applying several data mining tools such as RapidMiner, SPSS, WEKA, and Python. From the results of the comparison carried out resulted in 5 calculations, namely the manual calculation of cluster 1 with a ratio value of 73% being the first priority, calculations using RapidMiner resulting in cluster 3 with a ratio value of 58% being the first priority, calculations using SPSS cluster 2 with a ratio value of 34% being the first priority, and calculations using Python produce cluster 1 with a ratio value of 55% being the first priority.

Traffic congestion is a persistent challenge in urban areas in Indonesia, where increasing vehicle density creates the need for intelligent traffic monitoring systems. This study aims to develop a real-time vehicle parking system using the YOLOv8 object detection model to provide efficient traffic analysis from live CCTV broadcasts and recorded videos. This study uses a quantitative experimental approach with the implementation of the YOLOv8m model using the Ultralytics library in Python, tested on data collected from CCTV cameras A TCS Dishub Medan and additional footage from mobile devices. Vehicles are detected and counted in two directions up (Up) and down (Down) using virtual detection lines on the video frame. The system performance is evaluated by automatic detection counting with manually recorded ground truth data. The results show that on live CCTV broadcasts, the YOLOv8m model achieves an average precision of 98.96%, a recall of 96.59%, and an F1 score of 97.74% for upstream traffic, while for downstream traffic it achieves 100% precision, 95.64% recall, and an F1 score of 97.730/0. On the other hand, on high-quality recorded videos, all performance metrics achieve 100%, indicating perfect detection accuracy. These findings confirm the effectiveness of YOLOv8 in real-time traffic monitoring, but also indicate that video quality and stream stability affect detection performance. In conclusion, the developed system shows strong potential to support smart city traffic management solutions. Future research should focus on performance optimization under low-resolution live streaming conditions to improve accuracy in practical applications.  

Fully Homomorphic Encryption (FHE) enables computation on encrypted data with end-to-end confidentiality; however, its practical adoption remains limited by substantial computational costs, including long encryption and decryption times, high memory consumption, and operational latency. Zero-Knowledge Proofs (ZKPs) complement FHE by enabling correctness verification without revealing sensitive information, although they do not support encrypted computation independently. This study integrates both techniques to enable encrypted computation with verifiably consistent results. A prototype system is implemented in Python using Microsoft SEAL for homomorphic encryption and PySNARK for Zero-Knowledge Proof verification. Experiments are conducted on standard consumer-grade hardware (Intel i5, 8 GB RAM, Ubuntu 22.04) using datasets ranging from 100 MB to 1 GB. The evaluation focuses on encryption and decryption time, homomorphic computation latency, memory usage, and proof generation overhead. Experimental results show that integrating ZKPs introduces a moderate and stable runtime overhead of approximately 15–20%, as analyzed in Section 4, while enabling verification without plaintext disclosure. Ciphertext expansion remains a notable limitation, with observed growth of approximately 30–40× relative to plaintext size, consistent with prior FHE implementations. Despite these overheads, the system demonstrates feasible scalability for datasets up to 1 GB on mid-level hardware. Overall, the results indicate that the integrated FHE+ZKP approach provides a practical balance between confidentiality, verifiability, and performance, supporting its applicability to privacy-preserving scenarios such as secure cloud computation, encrypted data analytics, and confidential data processing under realistic resource constraints.

Polusi udara partikulat halus (PM2,5) merupakan ancaman serius bagi kesehatan masyarakat di Kabupaten Brebes, Jawa Tengah. Faktor penyumbang utamanya adalah emisi kendaraan di jalur Pantura, aktivitas industri perikanan, serta konsentrasi tinggi selama musim kemarau (Juni–November). Tidak adanya model peramalan sub-jam yang akurat menghambat pengembangan sistem peringatan dini yang efektif. Penelitian ini mengembangkan dan mengevaluasi model deep learning berbasis Transformer untuk memprediksi konsentrasi PM2,5 dengan resolusi waktu 15 menit. Data yang digunakan berasal dari NASA GEOS-CF (band PM25_RH35_GCC) yang diakses melalui Google Earth Engine menggunakan API Python. Dataset mencakup periode 1 Januari hingga 22 November 2025, menghasilkan 7.813 observasi per jam, yang kemudian diinterpolasi linear menjadi 31.249 titik data dengan resolusi 15 menit. Arsitektur Transformer terdiri dari 3 lapis enkoder, 4 kepala perhatian multi-head, dimensi embedding 128, dimensi feed-forward 256, panjang sekuen 60 timestep, dan augmentasi fitur menggunakan rerata bergulir (*rolling mean*, jendela = 3) dan beda pertama (*first difference*). Pelatihan dilakukan dengan TensorFlow-Keras, pengoptimal Adam, penjadwal peluruhan kosinus (*cosine decay scheduler*), dan fungsi kerugian Huber. Pembagian data dilakukan secara kronologis: 70% pelatihan, 30% validasi. Evaluasi pada set uji independen (16 Agustus–21 November 2025, 9.357 observasi atau 97 hari 11 jam 15 menit) menghasilkan MAE 0,7691 µg/m³, RMSE 1,2052 µg/m³, R² 0,9945, dan *Explained Variance Score* 0,9948. Model ini mampu menggambarkan variasi diurnal dan anomali musiman secara akurat, jauh melampaui model LSTM dan GTWR konvensional. Penelitian ini memberikan kontribusi signifikan di bidang Teknologi Informasi melalui kerangka kerja pengolahan *big data* satelit untuk aplikasi lingkungan.

The concept of supremum is fundamental in real analysis and plays a crucial role in the optimization of single-variable real functions. In practice, not all functions attain their supremum explicitly, which necessitates numerical approaches to evaluate their behavior computationally. This study aims to analyze the supremum of several one-dimensional real functions with different characteristics using a grid-search method implemented in Python. Four functions were examined: a parabolic function, a rational function with a sharp peak, a discontinuous piecewise function, and a function with a vertical asymptote. The analysis involved modeling the functions, discretizing the domain, performing numerical approximation of the supremum, verifying the results against analytical values, and using graphical visualization to observe the function behavior near the supremum. The findings indicate that the supremum of the parabolic, rational, and piecewise functions can be accurately identified, with results consistent with analytical expectations despite minor deviations caused by grid resolution limitations in the rational function. Meanwhile, the function with a vertical asymptote yields an unbounded supremum, which cannot be attained within the domain. These results demonstrate that Python provides stable and reliable numerical estimates of the supremum across various types of one-dimensional real functions, validating the effectiveness of computational methods in supporting conceptual understanding of supremum.

Hijaiyah letters have varying shapes, and some of them are very similar, often causing errors in the manual character recognition process. This study aims to classify Hijaiyah letters based on digital images using the Convolutional Neural Network (CNN) method. This method was used in this study with a dataset consisting of 28 letter classes and a total of 4,480 images obtained from various public sources and private data. All images underwent a preprocessing stage that included labeling, resizing, normalization, and augmentation, then were divided into three parts, namely training data, validation data, and test data with a ratio of 70:20:10. The training process was carried out using the Python programming language with the help of the TensorFlow and Keras libraries on the Google Colab platform. The test results showed that the CNN model achieved an accuracy of 97.10%, with an average precision, recall, and F1-score of 0.97, respectively. Classification errors only occurred in letters that had similar shapes, such as Syin and Sin. Based on these results, the CNN method proved to be effective, efficient, and accurate in recognizing Hijaiyah letter image patterns, so it can be used as a basis for developing classification models with higher accuracy in the future.

This study aims to simulate the M/M/1 queueing system using Python through a Modeling and Simulation approach supported by the Discrete-Event Simulation (DES) method. The objective of the research is to analyze key performance indicators of queueing behavior, including arrival time, service time, waiting time, queue length, and server utilization. The methodology employs DES, which models system behavior based on discrete events such as customer arrivals, service initiation, and service completion. The simulation generates stochastic arrival and service times using Poisson and exponential distributions, respectively. The results indicate that the DES-based M/M/1 simulation accurately reflects theoretical queueing behavior, showing increases in waiting times and queue lengths when arrival rates approach service rates, while server utilization corresponds to system load intensity. The findings demonstrate that DES is an effective approach for analyzing queue performance and can be extended to more complex models such as multi-server systems, priority queues, and predictive simulations using artificial intelligence.

Administrative services for civil servants (ASN) play a crucial role in ensuring the rights and obligations of government employees. One of these services is the submission process for the Wife Card (KARIS) and Husband Card (KARSU), which serve as official proof of marital status for civil servants. At the BKPSDM of Sumba Barat Daya, the current submission process for KARIS and KARSU is still carried out manually by filling out paper forms and attaching physical documents. This manual method is time-consuming, prone to data errors, and complicates the monitoring process. Therefore, this study aims to develop a digital information system to simplify the submission, verification, and issuance process of KARIS and KARSU online. This research applies a Research and Development (R&D) approach using the Waterfall model, consisting of requirement analysis, system design, implementation, testing, and evaluation stages. The system is developed using the Python (Streamlit) programming framework and a SQLite3 database, designed to be easily accessible to both ASN users and BKPSDM officers through local networks or the internet. The results show that the developed system successfully reduces processing time by up to 60%, improves data accuracy by 90%, and significantly decreases the use of physical documents. The implementation of this digital information system makes the KARIS and KARSU administrative service more efficient, transparent, and accountable, supporting the ongoing digital transformation within the local government of Sumba Barat Daya Regency.

MP3 audio files are often used in a variety of fields, but they are prone to security risks such as eavesdropping and illegal access. This study proposes a super encryption method by combining the algorithms of Rivest Shamir Adleman (RSA) and ElGamal to improve the protection of audio data. RSA was chosen for its efficiency and ease of implementation, while ElGamal offers a high level of security through discrete logarithmic complexity. The combination of the two is expected to address the weaknesses of each algorithm and strengthen file security. The system is developed using the Python programming language with the Visual Studio Code environment. The encryption process is done in layers: MP3 files are first encrypted with RSA, and then the results are encrypted again with ElGamal. The decryption is done in reverse order. Tests are conducted with various MP3 file sizes to measure the effectiveness and performance of the system. The results show that this method is able to secure audio files so that they cannot be accessed without the private keys of both algorithms. Although processing times are increased compared to single encryption, the level of security obtained is much higher. This approach can be an effective solution for protecting digital audio data, particularly MP3 format, and could potentially be applied to a wide range of other data types that require a high level of security.

Conventional attendance systems often face various problems such as inefficiency, inaccuracies in attendance logging, and limitations in recapitulation processes. Manual systems are prone to human error and time-consuming, while fingerprint-based systems may fail when the sensor is affected by dirty, wet, or damaged fingers. This study aims to develop an attendance system based on Artificial Intelligence (AI) by utilizing the face_recognition function in Python and implementing a microservice architecture to improve efficiency and accuracy in attendance recording. The system is developed using the Agile Feature-Driven Development (FDD) method, which focuses on building system features based on prioritized business values. This method is applied within the Software Development Life Cycle (SDLC) to ensure a structured, iterative, and user-oriented development process. Facial recognition is performed by comparing the encoding of the captured face image with the data stored in the database. The results show that the system is capable of recording attendance automatically, accurately, and in real-time. Furthermore, the recapitulation process becomes more efficient as the system manages and presents the data systematically.

Shallots are one of the most strategic horticultural commodities in Indonesia, with high demand and varying production levels across regions. Differences in productivity between areas often create challenges in managing distribution and formulating national food policies. This study aims to analyze shallot production data in Indonesia by applying the K-Means Clustering algorithm using Python. The production data were collected from official agricultural statistics publications, followed by preprocessing, normalization, and determination of the optimal number of clusters using the Elbow method and Silhouette Score. The clustering results show the formation of several groups representing regions with high, medium, and low production levels. Visualization of the clustering results reveals the distribution patterns of shallot production, which can serve as a basis for supporting policy formulation in the development of shallot production centers in Indonesia. Thus, the application of K-Means Clustering with Python proves to be an effective approach to provide clearer insights into regional production variations and can be utilized as an analytical tool to support decision-making in the agricultural sector.

Background: The scheduling of work of health workers, especially radiographers in type B hospitals, is a complex challenge due to the variety of radiology modalities, variations in the number of human resources, and the provisions of working hours regulations from the Ministry of Health of the Republic of Indonesia. Manual scheduling that is still in use tends to cause workload inequality, conflicts between employees, and operational inefficiencies. Objective: This study aims to design and develop an Artificial Intelligence (AI)-based radiographer shift scheduling system that is able to prepare work schedules automatically, fairly, flexibly, and integratedly, in accordance with hospital service regulations and needs. Research Method: This type of research is Research and Development (R&D). The development process is carried out through the stages of needs analysis, designing Python and Flask-based systems, simulating tests on data, and expert validation then the data collected and described from the initial mapping and also mapping potential problem-solving. Results: The system successfully manages morning, noon, night, and holiday shift schedules based on competence, fair rotation, and maximum working hours provisions. By showing a significant difference between user perceptions before and after using the system, which reflects improved efficiency, fairness, and ease of access to schedules. Respondents expressed satisfaction with the override feature and integrated notifications. Conclusion: The development of an AI-based radiographer shift scheduling system has proven to be feasible and effective in overcoming managerial problems of work scheduling in hospitals. This system is able to increase efficiency, transparency, and user satisfaction, and has the potential to be widely applied to various types of hospitals in Indonesia.

Digital image processing is a rapidly developing branch of computer science and has many applications in everyday life. One of the fields that most often utilizes this technique is object detection and color identification in images and videos. This study specifically aims to implement the thresholding method in the HSV (Hue, Saturation, Value) color space to detect three basic colors, namely red, green, and blue, in digital images. The research process begins with uploading images using the Google Colab platform, a cloud-based computing environment that makes it easy for users to run Python programs without requiring additional software installation. After the image is uploaded, the next step is to convert it from the RGB (Red, Green, Blue) color space to the HSV color space. This conversion is important because the HSV color space is more suitable for use in the color segmentation process. The Hue value represents the type of color, Saturation shows the level of saturation, while Value describes the level of brightness. Once the image is in the HSV color space, the next step is to determine the HSV value range for each basic color. This range is determined based on experimental results and references from related literature. Using this range, masking is performed to extract the appropriate pixels so that only the red, green, or blue portions of the image are visible, while the other colors are reduced. The results show that the thresholding method in the HSV color space is capable of detecting primary colors with a good level of visual accuracy, especially in simple images with contrasting backgrounds. The implementation of this program is relatively lightweight, easy to run directly in Google Colab, and does not require high-spec hardware. Therefore, this method is very suitable for use as basic learning material for digital image processing, both for students and novice researchers.

Employee shift scheduling in the hospitality industry remains a critical yet complex task due to fluctuating operational demands, fairness requirements, and labour regulations. Many hotels still rely on manual scheduling methods, which are time-consuming and prone to biases, particularly in ensuring fair workload distribution across employees. Despite numerous studies on workforce scheduling, limited attention has been given to integer linear programming (ILP) models that address gender-based restrictions and operational fairness simultaneously in real-world hotel contexts, especially in developing regions such as Central Java. This study proposes an Integer Linear Programming (ILP) model to generate optimal shift schedules for hotel staff over a 31-day planning horizon. The model incorporates operational constraints, including one shift per day, gender-based restrictions (which prevent female staff from working night shifts), availability, minimum staffing levels, and fairness in workload distribution. Key parameters and binary decision variables were defined to ensure compliance with the hotel's specific requirements. Empirical data were collected from a hotel in Central Java involving 20 employees, and the model was implemented using Python with a Gurobi solver. The ILP model successfully generated optimal schedules in under 10 seconds, significantly outperforming the manual method, which required over 4 hours. While the manual schedule resulted in an imbalance where some employees worked over 27 days and others only 22, the ILP approach enforced a strict maximum of 26 working days for all staff. Furthermore, the fairness index (FI) improved from 19.2% in the manual method to 0% in the ILP-generated schedule, indicating complete equity in workload allocation. The proposed ILP model demonstrates its effectiveness in improving scheduling fairness, operational efficiency, and compliance with labour policies. This work not only addresses a critical research gap in hospitality scheduling practices in Indonesia but also offers a replicable framework for other labour-intensive service sectors. Future research may explore multi-objective extensions incorporating employee preferences, satisfaction, and dynamic demand fluctuations.

Integral equations are essential tools in applied mathematics, with wide-ranging applications in fields such as physics, engineering, and finance. However, solving these equations presents significant challenges, particularly when dealing with complex, high-dimensional, or singular problems. Traditional methods, such as manual analytical techniques or direct numerical approaches, often struggle with computational efficiency, especially for large-scale systems, and may not be suitable for handling ill-conditioned problems. This study aims to develop an efficient numerical method for solving integral equations by combining adaptive quadrature techniques with Python-based iterative solvers. The adaptive quadrature method adjusts the step size dynamically based on error estimates, ensuring high accuracy even in the presence of singularities or near-singularities, which are common in many real-world problems. The iterative solver, based on Krylov subspace methods, enhances computational efficiency by reducing memory usage and improving the convergence speed of the solution. By using these techniques together, the proposed method significantly improves the computational time required to solve large-scale and complex systems of integral equations, while maintaining satisfactory accuracy. The results demonstrate that the adaptive quadrature technique, when combined with the Python-based iterative solver, offers a substantial advantage in both speed and precision compared to traditional methods. The proposed method is especially effective in handling complex, high-dimensional systems and ill-conditioned problems, making it a powerful tool for applied mathematics, physics, and engineering applications. In conclusion, this study presents a robust and efficient approach for solving integral equations, with potential for future research in solving non-linear and multi-dimensional integral equations.

Digital learning applications like HeyJapan are increasingly popular. User reviews on platforms such as Google Play Store contain valuable information on user perceptions and experiences. To process this information systematically, this study employs a Natural Language Processing (NLP) approach to analyze sentiment toward the HeyJapan application. Data was collected using web scraping techniques with Python and the google play scraper library, resulting in 1,000 latest user reviews. The analysis included data collection, preprocessing, sentiment labeling using TextBlob, visualization, modeling with Logistic Regression, and evaluation. After preprocessing, 923 valid reviews were classified into three sentiment categories based on polarity which are positive, neutral, and negative. Results showed 71.4% of reviews positive, 26.1% neutral, and 2.5% negative. Visualizations in pie charts and word clouds provided an overview of user perceptions. Modeling with TF-IDF and Logistic Regression achieved 88% accuracy with the highest f1-score in the positive sentiment category. Evaluation indicates the model is fairly reliable in classifying sentiments, especially for positive and neutral categories, though negative sentiment classification needs improvement. This study shows the NLP approach can evaluate user perceptions of educational applications based on reviews and serve as a basis for improving foreign language learning app quality.

The contemporary landscape of data management, marked by an unprecedented scale and velocity of data, has spurred the widespread adoption of NoSQL databases, prioritizing scalability and performance over traditional relational constraints. While offering significant flexibility, this paradigm shift introduces complex cybersecurity challenges, notably query injection vulnerabilities, which are consistently ranked among the top web application security risks. Redis, a leading in-memory key-value store powering critical infrastructure globally, presents a unique security profile due to its architectural design and features like Lua scripting. Despite its prevalence, a comprehensive academic evaluation of Redis injection attack vectors remains understudied. This study addresses this gap by systematically evaluating command and Lua script injection vulnerabilities in Redis version 7.4.1 across controlled configurations: default, password-protected, and ACL-secured environments. We quantify vulnerability risk and empirically validate mitigation strategies by employing a Dockerized testing framework, Python-driven exploit simulations, and CVSS v3.1 scoring. Our findings reveal critical weaknesses in default and permissively configured environments and demonstrate that restrictive Access Control Lists (ACLs), adhering to the principle of least privilege, provide complete mitigation against the specific injection vectors evaluated in our controlled experimental setup. We propose a Redis-specific threat taxonomy and provide empirically validated recommendations for securing Redis deployments, emphasizing layered security controls and proper ACL implementation. This research contributes the first systematic evaluation of modern Redis injection vulnerabilities and highlights the critical importance of security-conscious configurations to protect vital data infrastructure.