Journal of Information System Research (JOSH)

Implementasi Algoritma Kriptografi RSA untuk Keamanan Transmisi Data pada Sistem Monitoring Energi Listrik Berbasis IoT

Rajawali Rajawali — 2026-04-04

Data security is a crucial issue in Internet of Things (IoT) systems used to monitor electricity consumption. This study aims to enhance the security of data transmission in an IoT-based electricity monitoring system by implementing the Rivest–Shamir–Adleman (RSA) cryptographic algorithm. Data from the PZEM-004T sensor is encrypted using the RSA public key and verified with a digital signature before being transmitted to the server. The system was tested under two conditions: without encryption and with RSA encryption, including a simulated ARP spoofing attack using Ettercap. The results show that the system successfully rejected manipulated data, with a packet loss rate of 2.08%, which is categorized as “very good” based on the TIPHON standard, and achieved a throughput of approximately 9.88 bit/s. The implementation of RSA proved effective in maintaining data integrity and authenticity, thereby improving the reliability of the IoT-based electricity monitoring system.

Implementasi Metode SAW pada Sistem Seleksi Siswa Baru Berbasis Web

Nabil Ahyan Annakhief — 2026-04-04

New student admission is a crucial process in educational institution management because it determines the quality of accepted students. The Mathlaul Anwar Foundation offers several selection pathways: scholarships, report card grades, achievement pathways, and transfer pathways. Currently, the selection process is still conducted manually, resulting in various problems such as delays in data processing, potential calculation errors, lack of objectivity, and low transparency of selection results. This research aims to develop a web-based New Student Selection System using the Simple Additive Weighting (SAW) method as a decision support system to assist in the ranking process and determine student graduation objectively and measurably. The research methods used include observation, interviews, and documentation. The system development utilizes the Waterfall model, which consists of the stages of needs analysis, design, implementation, testing, and maintenance. The implementation results show that the system is able to reduce the selection process time from an average of 5 days to 2 days (a time efficiency of 60%). The process of calculating grades and ranking, which was previously done manually for approximately 120 minutes for 100 applicants, can be accelerated to approximately 15 minutes using the system (an efficiency increase of 87.5%). System testing using the Black Box method on 20 test scenarios showed a 100% functional success rate according to user requirements. In addition, the results of the SAW method calculation validation showed 100% accuracy compared to manual calculations. Thus, the application of the SAW method in the web-based new student selection system has been proven to be able to increase the efficiency, accuracy, objectivity, and transparency of the selection process at the Mathlaul Anwar Foundation.

Improving the POSPAY Mobile Interface Using User-Centered Approach with User Experience Questionnaire Evaluation

Tasya Arnomel Mareta — 2026-04-04

Digital public service applications require interfaces that are clear, efficient, and consistent to support fast and accurate transactions. In the PT Pos Indonesia service environment, POSPAY users may experience difficulties in locating core services, understanding menu structures, and completing tasks efficiently due to navigation and interface consistency issues. This study aims to improve the POSPAY mobile interface using a user-centered approach and to evaluate user experience using the User Experience Questionnaire. The study involved 20 participants (staff and customers). Observation and semi-structured interviews were conducted to elicit user needs, which were translated into prioritized requirements and implemented in a high-fidelity clickable prototype developed with Figma. Participants completed standardized task scenarios before completing the questionnaire. The results show positive mean scores in five dimensions, with Perspicuity (1.70) and Efficiency (1.55) as the highest, followed by Attractiveness (1.45), Dependability (1.20), and Stimulation (1.05). Novelty (0.65) remained neutral, indicating that the proposed interface is perceived as functional but not strongly innovative. The main contribution of this study is a context-specific requirement set and traceable mapping between user needs and prototype features for POSPAY in a postal service setting, supported by quantitative user experience evidence to prioritize interface refinement and implementation decisions at PT Pos Indonesia.

Evaluation of Service Quality Gaps in Pos Express Services Using the SERVQUAL Method

Muhamad Alif Fitrah Adriansyah — 2026-04-04

This study aims to analyze the service quality of Pos Express in South Sumatra by applying the SERVQUAL method to identify gaps between customer expectations and perceptions. A quantitative approach was employed by distributing structured questionnaires to 120 respondents selected through purposive sampling. The measurement instrument was developed based on five SERVQUAL dimensions: tangibles, reliability, responsiveness, assurance, and empathy. The results indicate that customer expectations were consistently higher than perceived service performance across all dimensions. The largest negative gap values were found in the responsiveness (-0.73) and reliability (-0.72) dimensions, indicating weaknesses in service response time, complaint handling, delivery punctuality, and information accuracy. Meanwhile, the empathy dimension recorded the smallest gap (-0.29), suggesting relatively positive interpersonal interactions between staff and customers. To support data processing and analysis, a web-based evaluation system was developed to automate SERVQUAL calculations and reporting. The system facilitated efficient data management and improved the accuracy of service quality analysis. Overall, the findings highlight the need for service improvement, particularly in enhancing operational reliability and responsiveness. This study provides empirical evidence to support service quality management and decision-making in regional postal services.

Penerapan Metode Association Rule Mining Menggunakan Algoritma Equivalence Class Transformation Dalam Menganalisis Pola Stok Obat

Aniq Astofa — 2026-04-04

Poorly planned drug inventory management often leads to imbalances between patient needs and the availability of medicines in clinics. This issue generally arises because transaction data has not been optimally utilized as a basis for decision-making. The purpose of this study is to identify patterns of drug associations by applying Association Rule Mining techniques using the Equivalence Class Transformation (ECLAT) algorithm. The research adopts a quantitative approach, utilizing one year of drug transaction data. The analysis reveals several combinations of medicines that are frequently prescribed together by healthcare providers. These association patterns provide valuable insights into prescribing tendencies within the clinic. By understanding the most common combinations, managers can plan drug procurement more accurately and efficiently. The information obtained not only helps anticipate the risk of stock shortages but also prevents excessive inventory that could result in waste. Thus, the application of the ECLAT algorithm proves effective in enhancing drug inventory management. Furthermore, the findings of this study can serve as a foundation for developing more efficient procurement strategies, ultimately improving the quality of healthcare services in clinics. Overall, leveraging transaction data through Association Rule Mining contributes significantly to evidence-based decision-making. This demonstrates that integrating data analysis techniques with inventory management can create a healthcare system that is more responsive, efficient, and patient-centered.

Implementasi Model Deep Learning MobileNetV2 untuk Klasifikasi Citra Melanoma Berbasis Web

Deva Safara Alfan — 2026-04-04

Melanoma is one of the most aggressive types of skin cancer with a high mortality rate if not detected at an early stage. In primary healthcare facilities, the lack of dermoscopy equipment causes examinations to rely solely on visual assessment, which may lead to diagnostic errors, particularly false negatives. This study aims to develop a web-based early melanoma detection system as a tool to assist initial screening. The proposed method implements a deep learning model based on the MobileNetV2 architecture using a transfer learning approach with pre-trained ImageNet weights. The dataset used in this study consists of melanoma and notmelanoma images from HAM10000, while the nonskin class is obtained from CIFAR-10 to help the model distinguish between skin lesion images and non-skin images. The dataset is divided into 70% training data, 20% validation data, and 10% testing data. Evaluation results show that the model achieves an accuracy of 90% in multiclass classification, while binary evaluation focusing on melanoma detection yields an accuracy of 90.48%, precision of 81.75%, recall of 91.96%, and an F1-score of 86.50% on the test data. The model is then implemented in a web-based system capable of displaying skin lesion classification results along with a confidence score in real time. The findings indicate that the developed system can perform automated image analysis and has the potential to be used as a supporting tool for early melanoma screening.

Pemanfaatan Algoritma FP-Growth pada Teknik Data Mining untuk Mengidentifikasi Pola Stok Produk Elektronik

Irawaty Irawaty — 2026-04-04

Managing the availability of electronic product stock is a crucial issue in the retail world due to the high variety of products and dynamic consumer purchasing patterns. Inaccuracy in determining the amount of stock can lead to excess inventory or product shortages, which impacts on decreasing operational efficiency. This study aims to apply the FP-Growth algorithm in the data mining process to determine the pattern of electronic product stock availability based on purchase transaction data. The dataset used in this study consists of 150 electronic product purchase transaction data. The main problem faced is the lack of optimal utilization of transaction data to determine the relationship between products that are frequently purchased together. As a solution, this study applies the Frequent Pattern Growth (FP-Growth) algorithm because of its ability to find association patterns without the need to generate candidate itemsets, making it more efficient in data processing. The research process begins with calculating the frequency of item occurrences, determining the minimum support value of 20% (30 transactions), forming an FP-Tree, and mining frequent itemsets and association rules. The results show that Mouse, Laptop, and Keyboard are the items with the highest frequency, respectively 80%, 73%, and 70% of the total transactions. The Mouse–Laptop–Keyboard purchasing pattern has a support value of 55% with a confidence level of 80%. While the Mouse → Keyboard rule yields the highest confidence level of 85%. Based on these results, it can be concluded that the FP-Growth algorithm is effective in identifying purchasing patterns for electronic products and can be used as a basis for decision-making in prioritizing stock availability more precisely and data-driven.

Pengelompokan Tanaman Perkebunan Berdasarkan Produktivitas dan Luas Lahan dengan K- Means Clustering

Ethaniel Williano Adhi Putra — 2026-04-07

Plantation data in West Java was grouped based on land area and crop productivity using the K-Means method. This data was obtained from Open Data Jabar from 2022 to 2024 and analyzed using a quantitative approach. Three groups can be identified based on the clustering results: one group has high productivity but relatively limited land area, another has large land area but suboptimal productivity, and the last group has equally low productivity and land area. The results indicate that land area does not always correlate with productivity. This study emphasizes the importance of selecting relevant variables and using methods consistently to produce more accurate and understandable analyses.

Sistem Pendukung Keputusan Penentuan Siswa Magang Terbaik Menggunakan Metode Simple Additive Weighting SAW

I Komang Sugiartha — 2026-04-11

Selecting the best interns is a crucial activity in assessing the success of the internship program at Budi Darma University. The manual assessment process often leads to subjectivity and is time-consuming. Therefore, a system capable of assisting in objective and efficient decision-making is needed. This study aims to develop a Decision Support System (DSS) for determining the best interns using the Simple Additive Weighting (SAW) method. The SAW method was chosen because it provides accurate results by summing the weighted scores of each alternative based on predetermined criteria. The assessment criteria used in this study include discipline, responsibility, communication skills, cooperation, and internship report results. Assessment data is processed by weighting each criterion, then calculated using the SAW formula to obtain each student's preference score. The results show that the system can assist the university in quickly and objectively determining the best interns. This system makes the assessment process more transparent, accurate, and supports data-driven decision-making.

Analisis Pengalaman Pengguna Light dan Dark Mode Pada Facebook dan Tokopedia Menggunakan Within-Subject Design

Muhammad Farhan — 2026-04-26

The use of light mode and dark mode in mobile applications is becoming increasingly common; however, their effects on user experience across different application contexts still need to be empirically examined. This study aims to analyze differences in user experience between light mode and dark mode in Facebook and Tokopedia. The study employed a quantitative approach using a within-subject design (repeated measures design) involving 25 respondents. Measurements were conducted using Time on Task, Error Rate, the User Experience Questionnaire (UEQ), and user preference. The results showed that Facebook had no significant difference in Time on Task, whereas Tokopedia showed significant differences in T1 (p = 0.024) and T3 (p = 0.047). For Error Rate, significant differences were found in Facebook T1 (p = 0.046) as well as Tokopedia T2 (p = 0.020) and T3 (p = 0.019). The UEQ results indicated that both modes were in the positive category without statistically significant differences. These findings suggest that the influence of display mode is more evident in specific task performance metrics than in overall user experience perception, indicating that its effects are contextual. This study contributes by providing a comprehensive evaluation that combines objective and subjective metrics to compare light mode and dark mode across two different application contexts.

Klasterisasi Siswa Berdasarkan Profil Akademik dan Karakteristik Belajar Menggunakan Algoritma K-Means untuk Mendukung Pembelajaran

Attaya Faiharani — 2026-04-26

Grouping students based on academic and non-academic characteristics is important to support the development of more targeted educational guidance strategies in schools. The main problem addressed in this study is the absence of objective data-based student mapping, which causes development programs to remain general and less targeted. This study aims to classify students using the K-Means clustering algorithm based on academic profiles and other supporting variables, and to evaluate cluster quality using the silhouette coefficient method. The research stages include data preprocessing, determining the optimal number of clusters, clustering using K-Means, and evaluating the clustering result. The results showed that four clusters were selected as the final configuration with a silhouette score of 0,1093, with cluster membership distributed into 12, 4, 2, and 2 students. Visualization using principal component analysis shows that most clusters are sufficiently well separeted. This study contributes a data-driven student grouping model that can be used as a basis for recommending student potential development according to the characteristics of each group.

Perbandingan K-Means dan DBSCAN dalam Analisis Pola Pergerakan Kapal Menggunakan Data Automatic Identification System (AIS)

Darmansah Darmansah — 2026-04-26

Batam waters are one of the busiest shipping lanes in Indonesia, with high ship traffic density and complex movement patterns. This condition requires data analysis techniques that can accurately identify and adapt ship movement patterns. The purpose of this study is to study ship movement patterns using Automatic Identification System (AIS) data, and also to see how the K-Means and DBSCAN algorithms work in the data clustering process. The AIS data used includes geographic coordinates, observation time, speed, and direction of ship movement in Batam waters. This study includes the application of the K-Means and DBSCAN algorithms, feature extraction and normalization, and data pre-processing to improve data quality. Internal validation metrics used to assess cluster quality are the Silhouette Score and the Davies–Bouldin Index. The results of the study show that the DBSCAN algorithm has a better level of cluster cohesion and separation between clusters than K-Means. The K-Means algorithm produces a Silhouette Score value of 0.48 and a Davies–Bouldin Index value of 0.91, while the DBSCAN algorithm produces a Silhouette Score value of 0.62 and a Davies–Bouldin Index value of 0.67. In addition, DBSCAN can find sound data of 19.96% of the data set, which indicates abnormal ship movements or does not form a certain density pattern. The results show that the DBSCAN algorithm analyzes ship movement patterns with AIS data in the Batam waters better than K-Means. This research is expected to be the basis for the development of maritime information systems that help monitor ship traffic, make decisions about safety, and manage waters.

Klasifikasi Siswa Berprestasi Berdasarkan Nilai Akademik dan Non-Akademik dengan Menggunakan Metode Random Forest

Ricky Gunawan — 2026-04-30

This study aims to develop a classification system for high-achieving students by integrating academic and non-academic aspects using the Random Forest method. The main problem faced by Natal State High School 1 is that the process of identifying high-achieving students still focuses on academic grades and does not yet comprehensively incorporate other indicators such as discipline, attendance, and extracurricular activities. This study employs a quantitative approach with data collection techniques including observation, interviews, and literature review. The data used were derived from the report cards for the odd-semester of the 2024/2025 academic year, covering 222 eleventh-grade students. The research stages included data preprocessing (data cleaning, transformation, normalization, and feature selection), data splitting using a stratified split (70% training data and 30% test data), and the application of the Random Forest algorithm for classification. The features used include average academic scores, absences (sick, excused, unexcused), and extracurricular activities. The results showed that the model performed very well, with an accuracy of 1.000 on the test data and an average cross-validation accuracy of 0.9865. Additionally, the precision, recall, and F1-score each reached 1.000. The classification results identified 13 students as high achievers, with the largest distribution coming from 11th grade class 1. These findings indicate that the Random Forest method is capable of producing accurate and consistent classifications and is effective in integrating various assessment indicators. This study is expected to support more objective and comprehensive decision-making within educational evaluation systems and to contribute to the development of more holistic classification models for assessing student success in school, based not only on academic achievement but also on important non-academic aspects.

E-Nutrition Label: Design and Architecture of a Web-Based Front-of-Pack Nutrition Labeling System

Yulita Sirinti Pongtambing — 2026-04-30

FOPNLs (Front of Pack Nutrition Labels) are nutritional labeling systems placed on the front of packaging to present nutritional information more simply. FOPNLs can help consumers quickly determine foods with better nutritional content and lower levels of salt, sugar, and fat. Nutrition labels influence consumer behavior and decision-making in determining healthy foods. However, the nutritional labeling system in Indonesia is not yet fully informative, and policies mandating that the food industry implement such labeling are not yet fully enforced. This study aims to develop an application model that automatically calculates FOPNLs for food products. The study resulted in a website prototype and limited testing, using the Design Science Research Method. DSRM can effectively bridge the theoretical foundations with practical requirements in the development of information system artifacts, particularly within the context of digital transformation in the healthcare sector. The result shows prototype functions well and can automatically calculate RDA and generate FOPNLs based on the nutritional label and serving size entered into the system. Functional evaluation using Black-Box Testing demonstrated a 100% success rate across all test scenarios, while the qualitative TAM-based assessment indicated that the proposed artifact was positively accepted, particularly regarding its perceived usefulness and ease of use. This prototype can be easily used by MSMEs that produce processed foods. Future research can be conducted through limited trials at the District or City UMKM Office.

Analisis Perbandingan Algortima Support Vector Machine, Random Forest dan Naive Bayes Untuk Prediksi Penyakit Kanker Paru-Paru

Rendy Alfa Rizky — 2026-05-09

The lungs are one of the vital organs responsible for the processes of respiration and blood circulation, with smoking habits being the primary factor contributing to the development of lung cancer. In Indonesia, the prevalence of this disease continues to increase, placing it eighth in the Southeast Asian region. Globally, lung cancer accounts for approximately 11.6% of all cancer cases and 18% of total cancer-related deaths.This study aims to analyze and compare the performance of Support Vector Machine (SVM), Random Forest, and Naïve Bayes algorithms in predicting lung cancer, as well as to determine the best-performing algorithm based on accuracy, precision, and recall metrics. The study utilizes the Lung Cancer Prediction dataset obtained from Kaggle, consisting of 309 instances and 16 attributes. The approach involves the implementation of three machine learning algorithms, namely Support Vector Machine (SVM), Random Forest, and Naïve Bayes. The research process includes data collection, preprocessing, data transformation, feature selection, model development, and evaluation using a confusion matrix. The experimental results show that both SVM and Naïve Bayes achieve the same accuracy of 91.07%, while Random Forest obtains an accuracy of 89.28%. In terms of evaluation metrics, SVM demonstrates more consistent performance with a precision of 95% and recall of 93%, whereas Naïve Bayes shows a higher recall of 95% with a precision of 93%. On the other hand, Random Forest exhibits limitations in identifying non-cancer cases. Based on the overall results, SVM is considered the most optimal method as it provides a better balance of performance. This study indicates that machine learning has significant potential as a supporting tool for early detection of lung cancer in a more accurate and efficient manner.

Sistem Informasi Manajemen Energi untuk Meningkatkan Efisiensi dan Ketepatan Pengelolaan Konsumsi Energi di Bandara

Daniel Arsa — 2026-05-09

Energy consumption recording (electricity, water, and fuel) at Sultan Thaha Airport Jambi currently relies on manual methods using spreadsheets. This process is prone to human error, data fragmentation, and hinders crucial energy efficiency analysis for airport operations. The need for data accuracy is becoming increasingly vital in line with the government's initiatives to tighten budget efficiency and drive digital transformation in public infrastructure. To support energy conservation policies and optimize operational costs, a more transparent and integrated monitoring system is required. This research aims to design and implement a web-based Energy Management Information System to address these challenges. Using the Waterfall development method, the system is built with a modern architecture utilizing React.js for a responsive interface, Express.js as the backend, and PostgreSQL for a scalable database management. Black Box testing results indicate that the system is valid and successfully provides an integrated solution through Dashboard Monitoring, Digital Input Validation, and centralized data reporting. The implementation of this system transforms energy governance from manual to digitally integrated, providing a solid foundation for airport management in making strategic decisions aligned with national budget efficiency and energy sustainability programs.