JOINS (Journal of Information System)

Application of the Naïve Bayes Algorithm for Classifying Graduation of Informatics and Computer Engineering Education Students at UIN SMDD Bukittinggi

2026-05-05T05:34:59+00:00

Timeliness of graduation is one of the indicators of university quality, and the utilisation of student data can provide valuable information to support decision-making. Quantitative data from the university's TIPD department, including gender, school of origin, Semester Grade Point Average (IPS), and Grade Point Average (GPA), are used as prediction attributes. Through the stages of data collection, attribute determination, data mining (cleaning, selection, transformation), and application of the Naive Bayes algorithm, a prediction model was built and tested. The results showed an accuracy of 87.5%, precision of 57.2%, and recall of 80%. It is concluded that the Naive Bayes algorithm is effective in classifying student graduation, with the funding source attribute identified as one of the influential factors. This study recommends the use of feature filtering such as information gain in future research to improve prediction accuracy.

Kimball Dimensional Modeling for Data Warehouse Design in a Manufacturing Enterprise

2026-05-05T13:49:44+00:00

Modern manufacturing companies face significant challenges from large data volumes and fragmented information systems, which hinder effective data-driven decision-making. This study aims to address these issues by designing a dimensional model for a data warehouse in an electronics manufacturing company, integrating scattered operational data into a single, unified repository. By applying Kimball's Business Dimensional Life Cycle methodology, this study systematically goes through four stages: defining core business processes, declaring data granularity, identifying dimensions, and identifying facts. The result is a fact constellation model (star schema) consisting of 11 grains, 8 fact tables, 8 star schema models, and 1 dimensional model. This proposed model simplifies data access for in-depth analysis, providing a robust and reusable framework to support strategic decision-making in a modern manufacturing environment

University Student Stress Detection Based on X Social Media Comments Using TF-IDF and Logistic Regression

2026-05-05T12:08:16+00:00

Mental health issues, particularly stress among university students, are on the rise and require special attention. Students tend to express their psychological conditions implicitly through comments or posts on social media, especially on platform X, which provides valuable digital data for real-time and non-invasive emotional analysis. This study aims to develop a stress detection system for students by analyzing comments on social media platform X using the Term Frequency-Inverse Document Frequency (TF-IDF) method and the Logistic Regression algorithm. TF-IDF is applied to extract important linguistic features from student comments, while Logistic Regression is chosen for its ability to provide clear probabilistic interpretation and efficiency in processing high-dimensional text data. The model is trained using labeled student comment data and evaluated using accuracy, F1-score, precision, and recall metrics. The results indicate that the system developed can classify stress and non-stress comments with a high accuracy of 93%, demonstrating great potential in supporting early interventions for student mental health. The implication of this research is expected to serve as a foundation for the development of digital applications that are responsive, adaptive, and practical in promoting student mental well-being in Indonesia.

Part-of-Speech Tagging in Javanese Using Pre-Trained Bidirectional Encoder Representation Model from Transformers

2026-05-05T12:44:07+00:00

Part-of-Speech Tagging (POS tagging) is the process of determining word classes in a text that is important in natural language processing. In Javanese, POS tagging is still a challenge due to limited linguistic resources and the complexity of the language. With the development of deep learning technology, the BERT (Bidirectional Encoder Representations from Transformers) fine-tuning method has been applied to classify word classes in Javanese, which is a language with limited resources. The javanese-bert-small model was trained using the UD_Javanese-CSUI dataset, and evaluated using precision, recall, F1-score, and accuracy metrics. The results showed that the model achieved good performance with an accuracy of 88,87%, and showed stability during training without significant overfitting. These findings indicate that the BERT-based approach is effective in handling word class ambiguity in Javanese and can be a stepping stone for further development in NLP systems for regional languages.

Cryptocurrency Trading Decision Support System Using Combination Weighting of EMA, RSI, MACD, and Bollinger Bands Indicators

2026-05-05T12:38:09+00:00

Cryptocurrency trading has rapid and significant price changes that cause investors to make decisions based solely on intuition when buying assets, potentially leading to a risk of loss. Therefore, this research aims to develop a cryptocurrency trading decision support system (DSS) using a combination of technical indicators, namely EMA, RSI, MACD, and Bollinger Bands. The system is designed to assist users in making more objective trading decisions based on historical data. This study applies weighted indicator combinations ranging from 0 to 4, resulting in 625 weight combinations evaluated thru backtesting using ROI, Win Rate, and MDD metrics. Based on the test results, the weighted indicator combination outperformed single indicators by achieving an ROI increase of up to 2222.35% on the SOLUSDT asset. In addition, the approach improved signal accuracy, as shown by the increase in Win Rate on ETHUSDT from 35.21% to 47.28% and on SOLUSDT from 32.84% to 58.11%. Furthermore, the method was effective in mitigating risk, indicated by the reduction of MDD on ETHUSDT from 50.04% to 41.35%. The system was successfully implemented as a web-based application integrated with Telegram notifications to deliver analysis results to users.

Decision Support System for Selecting Digital Platforms to Enhance MSME Product Brand Awareness Using the TOPSIS Method

2026-04-25T15:47:19+00:00

The development of digital technology has encouraged Micro, Small, and Medium Enterprises (MSMEs) to utilize digital platforms as a marketing tool to increase product brand awareness. However, the large number of digital platform options makes it difficult for MSMEs to determine the most effective platform. This study aims to develop a Decision Support System (DSS) in selecting digital platforms to increase MSME product brand awareness using the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) method. This study uses a quantitative approach by collecting data through questionnaires from MSMEs in Bogor Regency. The criteria used include customer reach, engagement level, promotion costs, ease of use, ease of features, and payment methods. The TOPSIS calculation results show that TikTok obtained the highest preference value of 0.657, followed by Shopee Marketplace at 0.574, Instagram at 0.473, Tokopedia at 0.441, and Facebook Pro at 0.392. These findings indicate that TikTok is the most recommended digital platform in increasing MSME brand awareness because it has a wide reach and high level of interaction. Thus, the implementation of TOPSIS-based DSS has been proven to be able to provide objective and systematic digital platform recommendations, thereby helping MSMEs formulate more effective and data-driven digital marketing strategies.

Development of an Inventory Management System with OCR and QR Code Integration for Student Identity Validation

2026-05-15T15:41:06+00:00

Inventory management at the Islamic Spiritual Activity Unit (UKKI) of UPN "Veteran" Jawa Timur still relies on manual methods, which are prone to human error and borrower identity manipulation. This study aims to develop a web-based inventory management information system integrating Optical Character Recognition (OCR) technology for identity validation and Quick Response (QR) Code for transaction security. Software development utilized the Rapid Application Development (RAD) method through an iterative approach with end-users. The results demonstrated that the OCR module, supported by pre-processing algorithms and Levenshtein Distance, successfully extracted and validated Student Identity Card (KTM) data automatically, achieving an acceptable similarity score (R) of over 0.87. Furthermore, implementing QR Codes as digital tokens proved effective in minimizing recording errors and ensuring officer accountability during handovers. Black-Box testing confirmed that 100% of the features, including dynamic stock management, operate precisely. In conclusion, this system successfully replaces conventional methods while enhancing time efficiency, data transparency, and organizational asset security.

Information System Audit on the Online Tax Pick-up Service of Tasikmalaya City SAMSAT Using the COBIT 4.1 Framework

2026-05-14T03:52:19+00:00

Digital transformation in the public service sector has encouraged the Tasikmalaya City Samsat to develop the Online Tax Collection Service in order to improve effectiveness, efficiency, and motor vehicle taxpayer compliance. The success of these digital services is influenced by the quality of information technology governance that supports operational activities. This study aims to evaluate the maturity level of information technology governance in the Online Tax Collection Service using the COBIT 4.1 framework. The research methods were conducted through observation, interviews, and questionnaires based on COBIT 4.1 domain indicators. The results showed that the average maturity level score was 2.73, approaching level 3 (Defined Process). The highest gaps were identified in AI6 (Manage Changes) and DS8 (Manage Service Desk and Incidents) because change management and incident handling processes were not formally documented. Based on the gap analysis, several recommendations were proposed, including the preparation of standard operating procedures, improvement of process documentation, implementation of change management, and development of human resource competencies. This study is expected to serve as a reference for improving information technology governance in regional digital tax services.

Sleep Disorder Detection Using Support Vector Machine on a Streamlit Web Application

2026-05-15T15:37:38+00:00

Sleep disorders are health problems that may affect an individual’s physical condition, mental well-being, and daily productivity. These conditions can be influenced by lifestyle and physiological factors, such as sleep duration, sleep quality, stress level, physical activity, heart rate, and blood pressure. This study aims to apply the Support Vector Machine (SVM) method to classify sleep disorders into three categories, namely normal, insomnia, and sleep apnea, as well as to develop a Streamlit-based web application to support interactive prediction. The dataset used in this study is the Sleep Health and Lifestyle dataset obtained from Kaggle. The research stages include data preprocessing, normalization using StandardScaler, model training using SVM and five comparison algorithms, and hyperparameter tuning to obtain the best performance. The evaluation results show that the SVM model with a poly kernel achieves an accuracy of 97.33% and a macro F1-score of 0.9569. The best model is then implemented into a web application that displays classification results along with the probability of each class, making it useful as an accessible early screening tool for sleep disorders.

Design of a Web-Based SISCA Information System for Monitoring OHS Emergency Equipment at PT Aisin Indonesia

2026-05-15T15:39:49+00:00

This study aims to design and develop a web-based System Information Safety Checksheet Aisin (SISCA) to support the digital, structured, and monitorable inspection process of occupational safety emergency equipment at PT Aisin Indonesia. The system development method applies the Systems Development Life Cycle (SDLC), consisting of planning, requirements analysis, design, implementation, testing, and maintenance. The system was developed using the Laravel framework with Model-View-Controller (MVC) architecture and MySQL database. The main features include role-based user authentication, QR Code scanning, inspection data entry, photo evidence upload, real-time monitoring dashboard, inspection history, and automatic report generation. The Black Box Testing results on 12 main functional scenarios show that all scenarios worked according to the requirements with a success rate of 100%. White Box Testing on the login, equipment inspection, and report generation modules resulted in cyclomatic complexity values of 3, 4, and 3, respectively. These results indicate that SISCA provides functions that meet user requirements, has simple program logic, and improves the effectiveness of inspection documentation, accelerates report access, and reduces the risk of human error.

Web-Based Lecture Attendance System with Dynamic QR Code Expired Per Session Using Prototype Method

2026-05-15T15:36:57+00:00

Manual student attendance recording in higher education still has weaknesses including risks of data manipulation such as proxy attendance, slow attendance reporting, and time inefficiency during lectures. This study aims to develop a web-based lecture attendance system utilizing dynamic QR Codes that are unique per session and automatically expire, accessible through smartphone browsers without additional application installation. The primary novelty lies in the automatic expired token mechanism, which renders QR Codes unusable after a session ends even if the code remains physically readable. The Prototype method was chosen for its suitability in rapid and iterative system development with active user involvement. The system was developed using PHP CodeIgniter 4, MySQL, phpqrcode library for QR Code generation, and jsQR for scanning via browser camera. Testing was conducted using Blackbox Testing on 12 functional scenarios and User Acceptance Testing (UAT) with a Likert scale to 20 respondents. All 12 Blackbox Testing scenarios succeeded with a 100% success rate. UAT results showed an acceptance rate of 89.71% categorized as Very Feasible. The system proved effective in improving the efficiency of real-time, accurate, and fraud-resistant student attendance recording.