Jurnal Kecerdasan Buatan dan Teknologi Informasi

WEBGIS APPLICATION FOR SEARCHING THE NEAREST CAR AC REPAIR SERVICE IN METRO CITY USING THE VINCENTY ALGORITHM

Rayhan Prasetiyo — 2026-05-02

The rapid increase in vehicle mobility in Metro City has led to a higher demand for specialized car AC maintenance services. This study develops a WebGIS-based application designed to facilitate the search for the nearest car AC workshops by implementing the Vincenty Algorithm. Unlike standard digital maps that often use spherical models, this system utilizes the WGS-84 ellipsoid reference to ensure high precision in geodesic distance calculations. The software was developed using the Waterfall model, integrating PHP, MySQL, and the Leaflet.js library. System validation was conducted through Black Box Testing across seven core modules, achieving a 100% functional validity rate. Comparative analysis between manual Vincenty calculations and system driving distance showed a minimal margin of 0.13 KM or 1.8%, confirming the algorithm's reliability for nearest-location ranking. This WebGIS serves as an efficient digital navigation tool to support vehicle maintenance for the community in Metro City.

AN ANALYSIS OF FEAR OF MISSING OUT (FOMO) AS A DRIVER OF HOAX DISSEMINATION IN THE PRABOWO ERA USING MLP

Irfan Saputra — 2026-05-02

The development of social media over the past decade has accelerated the spread of information, including hoaxes, which impact public perception and political stability. One psychological factor contributing to the impulsive spread of information is Fear of Missing Out (FOMO), defined as the feeling of anxiety experienced when individuals believe they are missing important information or events. This study aims to analyze the relationship between the FOMO phenomenon and the tendency to spread political hoaxes related to the Prabowo administration on social media. The research data was obtained through comment crawling techniques on the TikTok platform and then processed using the following stages: preprocessing text (e.g., cleaning, case folding, tokenizing, filtering, stemming) and labeling of FOMO, Non-FOMO, Hoax, and Non-Hoax classes. The Multi-Layer Perceptron (MLP) model is used to classify user behavior patterns. FOMO plays a role in increasing the spread of fake news in the political sphere, and this demonstrates that a combination of psychological factors and machine learning techniques can help understand the dynamics of disinformation on social media.

DESIGN AND DEVELOPMENT OF AN EARLY WARNING SYSTEM THROUGH CONTINUOUS AUDITING AND CONTINUOUS MONITORING IN PUBLIC SECTOR PROCUREMENT

R Wisnu Prio Pamungkas — 2026-05-02

The background of this research is the challenge in supervising goods and services procurement in the public sector, which is still dominated by traditional, reactive auditing methods conducted after transactions are finalized. The primary issue is the high volume of transactions and data complexity, which hinders early fraud detection. This research aims to design and develop an early warning system using Continuous Auditing and Continuous Monitoring (CACM) methods to enhance the effectiveness of fraud detection. The research method involves system development based on data integration from the Electronic Procurement System (SPSE) and other supporting monitoring systems. By utilizing data analytics, the system is designed to automatically identify risk indicators based on tender winner patterns and bidding behavior. The results indicate that CACM implementation enables real-time anomaly identification, providing early warning signals for auditors to take preventive measures before broader irregularities occur. In conclusion, the application of the CACM system transforms the internal oversight paradigm into a more proactive approach, strengthening fraud detection capabilities while improving accountability and transparency in government procurement processes

COMPARATIVE ANALYSIS OF RANDOM FOREST AND SUPPORT VECTOR MACHINE FOR FOOD CALORIE LEVEL CLASSIFICATION

Dading Oktaviadi Resmiranta — 2026-05-04

The rapid escalation of global metabolic health concerns emphasizes the critical urgency for advanced technological solutions that facilitate precise and automated monitoring of daily caloric intake. This research conducts a rigorous comparative analysis to evaluate the predictive performance and computational efficiency of Random Forest (RF) and Support Vector Machine (SVM) algorithms in classifying food calorie levels. The methodology commenced with a comprehensive data preprocessing phase involving multi-strategy missing value imputation and the discretization of caloric values into ordinal categories. Feature selection was meticulously executed using linear regression coefficients to identify high-impact nutritional variables. To ensure a robust evaluation, the dataset was partitioned using an 80:20 ratio for training and testing, complemented by cross-validation to minimize bias and variance. Experimental results indicated that the Random Forest (RF) demonstrated superior classification capabilities, achieving a peak accuracy of 94.8% alongside balanced precision and recall scores. Statistical evaluation via confusion matrices further revealed that Random Forest exhibited enhanced generalization across high-dimensional nutritional features compared to the geometric approach of Support Vector Machine (SVM). Furthermore, the analysis of computational overhead provided critical insights into the real-time deployment feasibility of each model. Ultimately, the findings suggest that the Random Forest serves as a robust engine for personalized dietary management systems, offering a reliable framework for future developments in preventive digital healthcare. By successfully bridging machine learning with nutritional science, this study establishes a benchmark for high-accuracy food classification essential for modern health-centric mobile applications.

SHORTEST ROUTE SEARCH TO ACCOMMODATIONS NEAR MANDALIKA CIRCUIT USING DIJKSTRA'S ALGORITHM AND ANDROID-BASED LOCATION-BASED SERVICE

Moch. Syahrir — 2026-05-04

The development of mobile technology, particularly on the Android platform, has created significant opportunities for real-time, location-based applications. One important implementation is the use of Location Based Service (LBS) in the tourism sector to help tourists efficiently find strategic locations. This study focuses on developing an Android-based LBS application that integrates the Dijkstra Algorithm to determine the shortest route to accommodations around the Mandalika Circuit area, Kuta Beach, Lombok, a leading destination for MotoGP events in Indonesia. The system development adopts the waterfall model, consisting of requirement analysis, system design, implementation, and testing. In the analysis phase, user needs related to accommodation information and route navigation are identified. The design phase includes system architecture, user interface, and digital map integration. Implementation is carried out by developing an Android application capable of accessing real-time location data and processing route calculations using the Dijkstra Algorithm to produce the most efficient path. The resulting application displays the distribution of nearby accommodations, provides travel distance information, and offers optimal route guidance that can be directly accessed by users. System testing shows that the application runs according to the defined functional requirements. Additionally, evaluation using a Likert-scale questionnaire indicates a user satisfaction level of 84%, reflecting good acceptance and usability. In conclusion, this research successfully implements LBS technology combined with the Dijkstra Algorithm in a mobile application, providing practical solutions for tourists visiting the Mandalika Circuit area.

WALMART PRICE PREDICTION USING HOLT-WINTERS FORECASTING

Melani Indriasari — 2026-05-04

Stock price prediction remains a complex challenge due to the volatile, noisy, and nonlinear nature of financial markets. This study aims to evaluate the effectiveness of the Holt-Winters Exponential Smoothing (HWES) method in forecasting the stock price of Walmart Inc. (WMT) and its application in investment decision-making. Historical monthly closing price data from January 2020 to December 2024 were collected and used to build an additive Holt-Winters model. The model was validated using out-of-sample data from January to February 2025, achieving RMSE of 4.535 USD and MAE of 4.801 USD, indicating good short-term predictive performance. The model was then used to forecast stock prices from March 2025 to December 2026, revealing a consistent upward trend with moderate seasonal fluctuations. However, deviations between predicted and actual values were observed during periods of market volatility, particularly in late 2025. To further evaluate performance, the Holt-Winters model was compared with the ARIMA model. Results show that ARIMA outperformed Holt-Winters in short-term forecasting with lower RMSE (4.71), MAE (4.26), and MAPE (4.21%), while Holt-Winters was more effective in capturing seasonal patterns. An investment simulation using a Dollar Cost Averaging (DCA) strategy combined with technical analysis indicators produced a total return of 3.45%, supported by both capital gains and dividend income. These findings suggest that while Holt-Winters provides a simple and interpretable approach for long-term forecasting, its performance can be improved by integrating adaptive models and external factors such as market sentiment and macroeconomic conditions for more robust predictions.

A SYSTEMATIC LITERATURE REVIEW ON THE INTEGRATION OF ARTIFICIAL INTELLIGENCE IN INFORMATION SYSTEM REQUIREMENTS ANALYSIS

Riski Akbar Matsuka — 2026-05-04

Requirements analysis is a critical phase in the development of information systems, as it significantly influences the overall success of a system. However, traditional approaches to requirements analysis are often performed manually and are prone to errors, inconsistencies, and inefficiencies. The advancement of Artificial Intelligence (AI) provides new opportunities to improve the effectiveness and automation of this process. This study aims to analyze the integration of AI in requirements analysis using a Systematic Literature Review (SLR) approach. The review follows the PRISMA 2020 guidelines and examines relevant studies published between 2020 and 2025. A total of 14 selected articles were analyzed to identify commonly used AI techniques, evaluate their effectiveness, and explore existing challenges. The results indicate that various AI techniques, including Machine Learning, Deep Learning, Transformer-based models, and Large Language Models (LLMs), have been widely applied in requirements analysis tasks such as classification, ambiguity detection, information extraction, and prioritization. These techniques demonstrate improvements in accuracy, time efficiency, and consistency compared to conventional methods. Despite these advantages, several challenges remain, including data imbalance, limited model generalization, lack of explainability, and limited validation in real-world industrial environments. Therefore, further research is needed to enhance the reliability and applicability of AI-based approaches in practical settings.

ENTERPRISE ARCHITECTURE PLANNING USING TOGAF ADM FOR FUEL DISTRIBUTION OPERATIONS

Putri Kaneshia Rahmadina — 2026-05-04

PT Dian Aristy Energi Palembang is a company engaged in the distribution of industrial fuel oil (BBM). The current operational processes are still conducted manually and are not integrated, leading to data duplication, reporting delays, low information accuracy, and difficulties in monitoring distribution activities, which affect managerial decision-making. This study aims to develop a strategic Enterprise Architecture plan based on TOGAF ADM to improve the alignment between information systems and fuel distribution operations. The research method used is qualitative descriptive with a case study approach, with data collection techniques including interviews, observations, and documentation. The TOGAF ADM phases applied consist of Preliminary Phase, Architecture Vision, Business Architecture, Information System Architecture, and Technology Architecture. The results of this study produce an Enterprise Architecture design that describes the current condition (AS-IS) and the proposed condition (TO-BE), including business process modeling, data architecture, application architecture, and supporting technology architecture. The proposed design enables the integration of operational processes through digital systems such as purchase order processing, distribution monitoring, and complaint management. This study concludes that the implementation of Enterprise Architecture based on TOGAF ADM can improve operational efficiency, data accuracy, information transparency, and support better decision-making, as well as provide a reference for the development of integrated information systems

EVALUATION OF IMBALANCE CLASS HANDLING STRATEGIES ON MACHINE LEARNING MODEL PERFORMANCE

Arry Verdian — 2026-05-05

Breast Cancer Dataset (BCD) represents a critical health problem due to the increasing prevalence of breast cancer and the importance of early detection of recurrence. Machine Learning (ML) approaches have been widely applied to support diagnosis and prediction; however, class imbalance remains a major challenge, where the majority class (“no-recurrence-events”) significantly outnumbers the minority class (“recurrence-events”). This imbalance can lead to biased models that fail to accurately detect recurrence cases. This study aims to evaluate the effectiveness of class imbalance handling using the Synthetic Minority Over-sampling Technique (SMOTE) on several ML models, including Decision Tree, Naïve Bayes, k-Nearest Neighbors (k-NN), and Random Forest. The dataset used consists of 286 records with 9 features obtained from the UCI Machine Learning repository. Data preprocessing was performed, including handling missing values and outliers, followed by class balancing using SMOTE. Model evaluation was conducted using 10-fold cross-validation and performance metrics such as accuracy, precision, recall, and F1-score. The results show that the application of SMOTE significantly improves model performance, with an average accuracy increase of 11.85%. Among the evaluated models, Random Forest combined with SMOTE achieved the best performance, with an accuracy of 79.79%. In contrast, models such as Naïve Bayes and k-NN demonstrated relatively lower performance. Overall, this study confirms that handling class imbalance using SMOTE can enhance classification performance, particularly in improving the detection of minority classes in breast cancer recurrence prediction tasks.

COMPARATIVE ANALYSIS OF PERFORMANCE OF MACHINE LEARNING FEATURE SELECTION IN EARLY DETECTION OF DIABETES

Lilik Joko Susanto — 2026-05-09

Diabetes is one of the most serious global health problems and continues to increase significantly worldwide. Early detection is essential to reduce complications and improve patient survival rates. Recently, Machine Learning (ML) has shown great potential in supporting early diabetes prediction through data-driven analysis. However, the presence of irrelevant and redundant features may decrease model performance and increase computational complexity. Therefore, this study aims to evaluate the effectiveness of feature selection techniques and ML algorithms for early diabetes detection using the PIMA Indians Diabetes Dataset. The dataset consists of 768 records, 8 features, and two classes. Data preprocessing was conducted to handle missing values and outliers using mean imputation and data cleaning techniques. Three feature selection methods were applied, namely Information Gain (IG), Gain Ratio (GR), and ANOVA, to identify the most relevant features. Furthermore, several ML algorithms, including k-Nearest Neighbor (k-NN), Random Forest, Support Vector Machine (SVM), Naive Bayes, and Neural Network, were evaluated using 10-fold cross-validation. The results showed that feature selection techniques improved classification performance compared to using all features. Glucose, BMI, Age, and Insulin were identified as the most influential features in diabetes prediction. Among all evaluated models, Random Forest combined with ANOVA achieved the best performance with an accuracy of 0.753. In general, the application of feature selection techniques increased model accuracy by up to 3.82%. These findings demonstrate that combining effective feature selection methods with robust ML algorithms can significantly enhance the performance of early diabetes detection systems.

VEHICLE REPAIR MONITORING INFORMATION SYSTEM FOR OPERATIONAL VEHICLES AT PT SU

Sultan Imam Fajri — 2026-05-09

Rapid developments in information technology have encouraged companies to improve operational efficiency through the implementation of integrated information systems. In transportation and logistics companies, vehicle maintenance management plays an important role in supporting operational continuity. PT SU currently still uses Microsoft Excel to record and monitor vehicle repairs, resulting in several problems such as data duplication, delays in reporting, difficulties in monitoring repair progress, and the risk of data loss. Therefore, this study aims to design and develop a web-based operational vehicle repair monitoring information system using the Web Engineering method. The development process consists of communication, planning, modeling, construction, and deployment stages. Unified Modeling Language (UML) was used to model system requirements, including use case diagrams and Entity Relationship Diagrams (ERD). The system was developed using PHP, MySQL, and Apache server through XAMPP. The developed system provides several features, including vehicle data management, repair requests, repair status monitoring, repair reports, and repair history management. System testing was conducted using black-box testing, performance testing, usability testing, and User Acceptance Testing (UAT). The testing results showed that all system functions operated properly according to user requirements. Performance testing indicated that the average response time was below 3 seconds, while usability testing showed positive results with ease of use reaching 90% and monitoring effectiveness reaching 92%. The developed system successfully improved repair data management, reporting efficiency, monitoring transparency, and coordination between departments at PT SU.

DESIGN OF AN ANDROID-BASED QUIZ GAME APPLICATION FOR INTRODUCING GCD AND LCM USING THE LCM METHOD

Robbi Gunawan — 2026-05-10

Learning Greatest Common Divisor (GCD) and Least Common Multiple (LCM) concepts often presents challenges for students due to conventional teaching methods that are less interactive and engaging. This study aims to design and develop an Android-based quiz game application for introducing GCD and LCM concepts using the Linear Congruential Method (LCM). The proposed application was developed as an interactive educational medium to improve students’ understanding and learning motivation through a game-based approach. The Linear Congruential Method was implemented to randomize quiz questions, ensuring varied question sequences and reducing repetition during gameplay. The application consists of several main features, including a home page, quiz gameplay, learning materials, high score tracking, and developer information. Additionally, immediate feedback mechanisms were integrated to indicate correct and incorrect answers, enabling students to learn from their mistakes directly. The implementation results show that the application successfully provides an interactive and engaging learning experience for students in understanding GCD and LCM concepts. Furthermore, the integration of question randomization using LCM contributes to creating a more dynamic learning process and increasing user engagement. Therefore, the developed application can serve as an alternative educational medium to support mathematics learning in a more effective and enjoyable manner.

LITERATURE ANALYSIS ON THE ROLE OF ARTIFICIAL INTELLIGENCE IN STRENGTHENING CYBERSECURITY IN E-GOVERNMENT SERVICES

Erfan Wahyudi — 2026-05-10

The rapid expansion of e-government services has increased the importance of cybersecurity in protecting public digital infrastructure, citizen data, and the continuity of government operations. In this context, Artificial Intelligence (AI) has emerged as a promising approach to strengthening cyber defense through real-time monitoring, anomaly detection, intelligent classification, and adaptive threat response. This study examines the role of AI in strengthening cybersecurity in e-government services through a Systematic Literature Review (SLR) of 27 selected articles published between 2019 and 2025. The review synthesizes the literature at the intersection of AI, cybersecurity, and digital government to identify major research trends, dominant methodological approaches, thematic classifications, and key implementation challenges. The findings show that AI is increasingly positioned not only as a tool for improving administrative efficiency, but also as a strategic enabler of cyber resilience in public-sector digital ecosystems. The literature highlights that machine learning, deep learning, explainable AI, anomaly detection, and privacy-preserving learning models have substantial potential for improving the security of citizen portals, digital identity systems, inter-agency platforms, and smart-government infrastructures. However, implementation remains constrained by fragmented data environments, interoperability problems, institutional readiness gaps, limited explainability, privacy concerns, and the dual-use nature of AI in cyber defense and cyber offense. This study concludes that AI is most effective when integrated into a broader socio-technical framework encompassing governance, accountability, transparency, and organizational capacity.

COLONOSCOPIC POLYP SEGMENTATION USING SEGFORMER-B0 WITH A DICE-BCE HYBRID LOSS

Ahmad Yani — 2026-05-10

Colorectal cancer is one of the leading causes of cancer-related deaths worldwide, with most cases originating from early lesions such as colon polyps. Early detection through colonoscopy is essential to reduce mortality rates; however, accurate polyp identification remains challenging due to variations in shape, size, texture, and illumination conditions. This study aims to implement and evaluate the SegFormer-B0 architecture combined with a Dice-BCE hybrid loss function for polyp segmentation in colonoscopy images. The study utilized the public Kvasir-SEG dataset consisting of 1,000 colonoscopy images with pixel-level annotations. The dataset was divided into 80% training data and 20% validation data. Image preprocessing included resizing to 256×256 pixels and normalization using ImageNet statistics. The model was trained for 25 epochs using the AdamW optimizer with a learning rate of 1×10⁻⁴. Performance evaluation was conducted using Dice Coefficient, Intersection over Union (IoU), Sensitivity, and Specificity metrics. The experimental results demonstrated that the proposed model achieved a Dice Coefficient of 89.92%, Mean IoU of 81.90%, Sensitivity of 89.12%, and Specificity of 98.51%. The training process also showed stable convergence, supported by a training loss of 7.53% and validation loss of 23.30%. The findings indicate that the integration of SegFormer-B0 with the Dice-BCE hybrid loss effectively improves segmentation accuracy and stability while addressing class imbalance issues in colonoscopy images. Therefore, the proposed approach has strong potential to support computer-aided diagnosis systems for colorectal cancer screening.

IMPROVING OPERATIONAL EFFICIENCY VIA END USER DEVELOPMENT: A WEB-BASED SALES MANAGEMENT SYSTEM FOR DR. BARON POMADE

M. Daud Mursal Lubis — 2026-05-10

This study aims to design and implement a Web -based sales management information system for Dr. Baron Pomade using the end user development (eud) method. The background of this research lies in sales processes that were previously conducted manually, which resulted in limited customer reach, slow transaction recording, and the absence of structured sales reports. The main objective is to develop a sales system that enhances operational efficiency, simplifies transaction management, and provides accurate information to support decision-making. The significance of this research is to offer a technological solution that replaces the manual process with an integrated digital system. Furthermore, it encourages active user participation in the system development process through the EUD approach, thereby producing an application that aligns with real operational needs. This study is also expected to serve as a reference for the development of WEB -based information systems in small and medium enterprises, particularly within the men’s cosmetic industry The findings demonstrate that the developed sales information system has been successfully implemented with key features including login, registration, product management, category management, transaction handling, reporting, and configuration settings. System testing confirmed that all features function as intended. The implementation of this system significantly improves business process efficiency, accelerates transaction recording, enhances data accuracy, and supports real-time decision-making. Thus, the study proves that applying the EUD method in sales information system development positively impacts the operational performance of the business.

Keywords: Sales Information System, End User Development (Eud), Pomade .

ENHANCING ONLINE AUTOMOTIVE SPARE PARTS SALES THROUGH A WEB-BASED E-SALES SYSTEM USING RAPID APPLICATION DEVELOPMENT

Muhammad Alif Fiqri Harahap — 2026-05-11

The rapid advancement of information and communication technology has encouraged business sectors to adopt internet-based systems to improve operational efficiency and competitiveness. In the automotive industry, particularly in the sale of used car spare parts, electronic sales (e-sales) have become an important solution for expanding market reach and facilitating transactions. However, this sector still faces several challenges, including limited product transparency, inconsistent stock management, unclear product conditions, and low consumer trust in online transactions. This study aims to develop a web-based e-sales system for used car spare parts using the Rapid Application Development (RAD) method to improve transaction efficiency, inventory management, and customer trust. The RAD approach was selected because it emphasizes rapid and iterative system development through continuous prototyping and active user involvement, allowing applications to be developed according to user needs in a shorter time. The research method consisted of four stages: requirements planning, user design, construction, and implementation. The resulting system integrates various functionalities, including product management, automatic inventory monitoring, customer registration and login, transaction processing, shipment tracking, website settings, and sales monitoring through an admin dashboard. The implementation results indicate that the developed system effectively improves operational efficiency by reducing manual recording activities and simplifying transaction management. Additionally, the system enhances transparency through detailed product information and transaction monitoring features, thereby increasing consumer confidence in purchasing used spare parts online. Overall, the implementation of the RAD method proved effective in developing an adaptive and efficient e-sales platform that supports digital transformation in the automotive spare parts.

COMPARATIVE STUDY OF CLASSIFICATION MODELS IN PROCESSING STUDENT TEST SCORES DATASETS

Rico Pramestiawan — 2026-05-11

The development of Machine Learning (ML) has contributed significantly to the field of education, particularly in analyzing student academic data to support data-driven decision-making. Predicting student exam results is important for identifying academic performance patterns, detecting potential failures, and improving learning interventions. However, variations in student characteristics and dataset complexity require the selection of appropriate classification models to achieve optimal prediction performance. This study aims to compare the effectiveness of several ML classification models in predicting student exam results using a student academic dataset. The dataset consists of 306 records, seven attributes, and five grade classes (A, B, C, D, and E), including attendance, quiz scores, midterm examination scores, final examination scores, and assignment scores. Data preprocessing was conducted to handle missing values, duplication, inconsistencies, and outliers. The dataset was split into training and testing data with a ratio of 75:25 and evaluated using 10-fold cross-validation. Several classification models were applied, including k-Nearest Neighbour (kNN), Decision Tree, Naive Bayes, Support Vector Machine (SVM), and Random Forest. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics. The experimental results showed that Random Forest achieved the best performance with an accuracy of 73.9%, precision of 74.0%, recall of 73.9%, and F1-score of 73.9%, followed by Naive Bayes and Decision Tree. Meanwhile, SVM produced the lowest performance among the tested models. The findings indicate that Random Forest is the most effective method for predicting student exam results and has strong potential to support educational decision-making systems.

COMPARATIVE ANALYSIS OF PERFORMANCE OF MACHINE LEARNING FEATURE SELECTION (GINI DECREASE AND RELIEF-F) IN HEART DISEASE DATASET

Chindu Lintang Bhuana — 2026-05-12

Heart disease remains one of the leading causes of mortality worldwide and presents a major challenge in healthcare systems. Early detection plays an essential role in improving survival rates and minimizing complications through timely intervention. Recent advances in Machine Learning (ML) have provided new opportunities for developing accurate and efficient prediction systems for heart disease detection. However, one of the major challenges in ML-based prediction is identifying the most relevant features to improve classification performance while reducing computational complexity and noise. This study aims to evaluate the effectiveness of two feature selection techniques, namely Gini Decrease (GD) and ReliefF, combined with several ML models, including Support Vector Machine (SVM), Tree, Naïve Bayes, and Random Forest, for heart disease classification. The study employed the UCI Heart Disease Dataset consisting of 303 records and 14 attributes. Data preprocessing included handling missing values using mean imputation, followed by feature selection and classification using 10-fold cross-validation with an 80:20 training-testing ratio. Experimental results showed that ReliefF outperformed GD, achieving the highest average accuracy of 0.796, compared to GD with 0.767 and all features with 0.771. The SVM model achieved the highest accuracy using GD (0.833), while Random Forest demonstrated optimal performance with ReliefF (0.817). Furthermore, the Tree model exhibited the fastest computational time among all evaluated models. These findings indicate that integrating suitable feature selection methods with ML models significantly enhances heart disease classification performance, particularly in improving predictive accuracy and computational efficiency for early medical diagnosis applications.

XGBOOST-BASED FRAUD TRANSACTION CLASSIFICATION ANALYSIS IN ONLINE PAYMENT SYSTEMS

Sri Diantika — 2026-05-14

The rapid development of online payment systems has significantly facilitated digital transactions; however, it has simultaneously increased the risk of fraudulent activities. Fraud detection has become a critical challenge due to the complex characteristics of transaction data and the imbalanced class distribution between legitimate and fraudulent transactions. This study aims to analyze the performance of the XGBoost algorithm in classifying fraudulent transactions within online payment systems. The research employs the Online Payments Fraud Detection Dataset obtained from the Kaggle platform. The research methodology consists of several stages, including dataset collection, data preprocessing, categorical data transformation using label encoding, feature engineering for the generation of new attributes, data partitioning through split validation with an 80:20 ratio, model development using the XGBoost algorithm, and performance evaluation using a confusion matrix, accuracy, precision, recall, F1-score, and Area Under the Curve (AUC). The experimental results demonstrate that the XGBoost model achieves excellent classification performance, with an accuracy of 99.98%, precision of 85%, recall of 100%, F1-score of 92%, and an AUC value of 0.9996. Furthermore, feature importance analysis reveals that errorOrig and newbalanceOrig are the most influential attributes in detecting fraudulent transactions. Based on these findings, it can be concluded that the XGBoost algorithm is highly effective for fraud transaction classification in online payment systems and exhibits strong potential for implementation in automated fraud detection systems to enhance the security of digital financial transactions.

DATA-DRIVEN CONSUMER SEGMENTATION APPROACH FOR JEANS RETAIL SALES USING FUZZY C-MEANS CLUSTERING

Nana Suarna — 2026-05-16

The fashion retail industry generates large volumes of sales transaction data containing valuable information regarding consumer purchasing behavior and preferences. However, extracting meaningful insights from heterogeneous retail data remains challenging when using conventional analytical approaches. This study aims to analyze jeans sales transaction data and identify consumer purchasing patterns using the Fuzzy C-Means (FCM) clustering algorithm. The proposed approach adopts the Knowledge Discovery in Databases (KDD) framework, consisting of data selection, preprocessing, transformation, data mining, and evaluation stages to ensure systematic analysis. The dataset used in this study consists of 799 jeans sales transaction records collected in 2024 from Shakila Collection, involving four attributes: product name, payment method, price, and purchase quantity. To improve clustering effectiveness, only price and purchase quantity were selected as the primary variables due to their relevance in representing consumer purchasing behavior. Clustering performance was evaluated using the Davies-Bouldin Index (DBI) to determine the optimal number of clusters. Experimental results show that the best clustering configuration was achieved at , producing three consumer segments consisting of 175 items in Cluster 0, 590 items in Cluster 1, and 34 items in Cluster 2. The findings indicate that medium-priced products tend to have higher purchasing intensity and more flexible purchase quantities, whereas premium-priced products exhibit relatively lower demand. The novelty of this study lies in integrating Fuzzy C-Means clustering with consumer preference analysis to generate practical business insights for pricing strategies, inventory optimization, and targeted marketing, thereby supporting more effective data-driven decision-making in fashion retail businesses.

USER INTERFACE DESIGN OF E-COMMERCE WEBSITES FOR MICRO, SMALL, AND MEDIUM ENTERPRISES (MSMEs) IN THE CULINARY INDUSTRY

Malisa Putri — 2026-05-18

Micro, Small, and Medium Enterprises (MSMEs) in the culinary sector play a strategic role in supporting economic growth, increasing employment opportunities, and strengthening local economies. However, many culinary MSMEs still experience challenges in adopting digital technology, including limited digital literacy, inadequate technological infrastructure, and manual sales and transaction management. These limitations hinder market expansion and reduce operational efficiency, making digital transformation increasingly necessary. This study aims to design a responsive and user-friendly user interface for a web-based e-commerce platform specifically intended for culinary MSMEs. The research method involved requirement analysis through observation, literature review, user needs identification, product data collection, order workflow analysis, and transaction management analysis. The interface design process was carried out using Figma to develop a prototype representing the overall system workflow, including customer and administrator interactions. The resulting design includes several key features such as the home page, login and registration page, product menu, payment system, order history, best seller recommendations, contact information, and logout functionality, as well as administrative features for product and transaction management. Black Box Testing was conducted to evaluate the functionality of each feature and ensure compliance with system requirements. The testing results demonstrated that all system functionalities operated successfully and consistently according to expected outcomes. The developed interface design is expected to support culinary MSMEs in improving digital marketing activities, simplifying transaction management, and increasing operational efficiency. Therefore, the proposed web-based e-commerce interface has strong potential to support the digital transformation and sustainability of culinary MSMEs in Indonesia.

DESIGN OF A PRELOVED MARKETPLACE WEBSITE APPLICATION FOR UNIVERSITY STUDENTS

Alfi Mahgfiro — 2026-05-18

The rapid development of information and communication technology has significantly influenced commercial activities, including buying and selling transactions conducted through digital platforms. Among university students, the increasing demand for affordable products has encouraged the growth of preloved or second-hand goods trading. However, existing general marketplace platforms do not specifically accommodate the needs of university students in conducting preloved transactions within campus environments, resulting in inefficiencies in product searching, limited transaction relevance, and concerns regarding security and trust. This study aims to design a website-based Preloved Marketplace application specifically intended for university students to facilitate secure, efficient, and affordable buying and selling activities for second-hand usable goods. The research employed a qualitative approach through observation, interviews, surveys, literature review, and system requirement analysis. The design process included problem identification, UI/UX design using Figma, prototype implementation, and system evaluation. The developed application provides various features, including user registration, login, product listings, product search and filtering, shopping cart, checkout, payment confirmation, live chat, sales reports, and dashboard monitoring. Black Box Testing was conducted to evaluate system functionality based on user input and output behavior. The testing results indicate that all system features operated successfully according to expected requirements, demonstrating functional consistency and usability. The developed platform is expected to support university students in selling unused but still usable products while helping other students obtain affordable necessities. Therefore, the proposed Preloved Marketplace has strong potential to improve transaction efficiency, support sustainable consumption, and encourage digital transformation within campus communities.

INVESTIGATING COMMUNITY READINESS THROUGH IT INFRASTRUCTURE, ONLINE TRANSACTIONS, AND COMMUNITY BEHAVIOR FOR URBAN VILLAGE DIGITAL TRANSFORMATION IN PALEMBANG

Berliana Inasti — 2026-05-20

The rapid development of information technology has encouraged digital transformation in various sectors, including public administrative services at the urban village level. However, the successful implementation of digital transformation depends not only on government readiness and technological infrastructure but also on the readiness of the community as the primary users of digital services. This study aims to evaluate community readiness toward digital transformation and design a web-based digital administrative service architecture for Sukamaju Urban Village, Palembang City, using the TOGAF ADM framework. The study employed a descriptive quantitative approach involving 395 respondents selected using the Slovin formula with a 5% margin of error. Data were collected through closed-ended questionnaires based on a 5-point Likert scale and analyzed using the E-Readiness approach through three main variables: Information Technology Infrastructure, Online Transactions, and Community Behavior. The results indicate that the Information Technology Infrastructure variable achieved the highest mean score of 4.32 (Very Ready), followed by Community Behavior with a mean score of 4.20 (Ready), and Online Transactions with a mean score of 4.11 (Ready). These findings suggest that the Sukamaju Urban Village community possesses a high level of readiness to support digital transformation in administrative services. Based on these findings, a web-based digital administrative service architecture was proposed using TOGAF ADM phases, including Architecture Vision, Business Architecture, Information Systems Architecture, and Technology Architecture. The proposed system is expected to improve service efficiency, reduce manual administrative processes, and support sustainable digital transformation at the urban village level.