Jurnal Techno Nusa Mandiri

OPTIMIZATION OF MACHINE LEARNING ALGORITHMS IN THE CLASSIFICATION OF VECTOR-BORNE DISEASES

2025-07-21T05:22:57+00:00

Developing a predictive model is the objective of this study, focusing on vector-borne diseases using various machine learning methods, including Random Forest (RF), Logistic Regression (LR), k-nearest Neighbors (kNN), Tree (DT), and XGBoost. The main goal is to use oversampling techniques like SMOTE and Random Oversampling to correct the dataset's class imbalance. The dataset was obtained from Kaggle and literature references published in Frontiers in Ecology and Evolution (Endo and Amarasekare 2022), consisting of approximately 9,490 entries with environmental, demographic, and clinical attributes. Dengue Fever is one of the diseases that this study focuses on. Aedes aegypti mosquitoes spread it, and it is a significant health risk in tropical areas. The DT and XGBoost models had the highest accuracy, at 99.2%. Logistic Regression and Random Forest also did well, with 99.1% accuracy. KNN did well, too, but with a lower recall, at 99.0%. The ROC curve gave a complete picture of how well each model classified things. These findings indicate that when combined with proper data handling, machine learning models can significantly improve early detection of vector-borne diseases and support more accurate and timely decision-making in public health interventions.

DESIGN OF WEB-BASED CAR RENTAL INFORMATION SYSTEM USING EXTREME PROGRAMMING AT CV. NUGROHO

2025-04-14T04:08:08+00:00

This study discusses the development of a web-based car rental information system for CV. Nugroho Trans Surabaya using the Extreme Programming (XP) methodology. The system was designed to address issues in the previous manual rental process, such as transaction recording on paper, which was prone to errors and delays in data management. The developed system includes key features such as car booking, fleet data management, rental confirmation, and payment integration. Testing was conducted through various methods, including performance testing, usability testing, and security testing. Performance testing using PageSpeed Insights in desktop mode showed the following scores: performance 93, accessibility 84, best practices 93, and SEO 82. Meanwhile, testing with GTmetrix yielded a performance score of 96%, a structure score of 72%, a Largest Contentful Paint (LCP) time of 909 ms, and a fully interactive time of 1.3 seconds, indicating excellent speed and interface stability. In terms of security, testing with Pentest Tools indicated an overall medium risk level, with 1 medium risk finding, 5 low-risk findings, and 13 informational findings, and no high-risk vulnerabilities. The application of the XP method enabled adaptive system development tailored to user needs and iterative changes. This system has proven to increase the company’s operational efficiency by up to 40%, based on faster transaction completion times compared to the manual system. However, some limitations remain, such as user interface constraints and suboptimal integration of online payment channels. For future research, it is recommended to improve user experience, optimize the mobile interface, enhance server security protection, and expand system features to support the broader business growth of CV. Nugroho Trans Surabaya

A LIGHTWEIGHT AND PRACTICAL PIPELINE FOR CROSS-PROJECT DEFECT PREDICTION USING METRIC-BASED LEARNING

2025-08-08T05:54:56+00:00

Cross-Project Defect Prediction (CPDP) addresses the scarcity of defect data in new software projects by transferring knowledge from existing ones. However, domain shift between projects remains a major challenge. This study introduces a lightweight and practical CPDP pipeline based on traditional metric features, integrating domain adaptation (CORAL, TCA, TCA+), feature selection, and resampling techniques. Through 120 configurations evaluated on multiple PROMISE datasets, we found that combining TCA or TCA+ with Synthetic Minority Over-sampling Technique combined with Edited Nearest Neighbors (SMOTEENN) consistently improved F1-Score and Recall on imbalanced datasets. LightGBM demonstrated the most stable performance across projects, while Logistic Regression yielded the highest MCC in specific cases. Principal Component Analysis (PCA) visualizations supported the effectiveness of domain alignment. The proposed pipeline offers a reproducible, cost-efficient alternative to deep learning models and provides actionable insights for defect prediction in resource-constrained environments.

LAND COVER CHANGE PREDICTION USING CELLULAR AUTOMATA AND MARKOV CHAIN MODELS

2025-09-08T02:43:33+00:00

This research examines the impact of land use change on mobility. Spatial problems arise due to increased activity, population, and transportation in the same space, necessitating the development of modeling strategies. This aligns with SDG 11 on cities and settlements, as well as the PRN's focus on transportation innovation. The urgency of this research lies in its adaptive and sustainable spatial prediction efforts aimed at controlling future land use. This study aims to analyze land use change patterns using the Cellular Automata Markov Chain (CA-Markov) model in Kupang City until 2043. CA-Markov simulations efficiently evaluate land cover changes and movement. The quantitative research method was conducted based on spatial predictions and spatial configuration. Quantum GIS (QGIS) and GeoSOS-FLUS were used to obtain results from each stage. There are three research stages. First, identification of land cover (land use in 2018 and 2023), driving factors (distance to settlements, airports, highways, elevation, slope, slope orientation, rainfall, population density), and conservation areas. Second, standardisation of spatial data. Third, land cover prediction using GeoSOS software (five-year prediction) to identify patterns of land use change. These findings emphasize the importance of using CA-Markov-based spatial predictions as a foundation for adaptive spatial planning to control land-use conversion and maintain sustainable spatial connectivity in Kupang City until 2043.

SENTIMENT ANALYSIS OF PUBLIC OPINION ON TRANSPORTATION SERVICES IN INDONESIA USING MACHINE LEARNING

2025-03-17T01:19:51+00:00

This study analyzes public sentiment towards transportation services in Indonesia through social media using Naïve Bayes and Support Vector Machine (SVM) algorithms. Data was collected from Twitter using an API with transportation-related keywords over a three-month period. The analysis results indicate that 93.5% of the opinions are neutral, 3.5% are positive, and 3% are negative. The dominance of neutral sentiment suggests potential dataset imbalance or user hesitation in expressing strong opinions. SVM achieved a higher accuracy (100%) compared to Naïve Bayes (92%), which may be influenced by dataset limitations or the model's validation method. Data preprocessing involved several steps, including tokenization, stopword removal, stemming, lemmatization, and handling of missing data to ensure cleaner and more structured text input. These findings highlight the potential of sentiment analysis for transportation policy improvements, providing insights for policymakers and transport service providers. Future research should address data balancing and broader dataset usage to enhance the robustness of findings and support better decision-making in the transportation sector.

ANALYSIS OF CUSTOMER SATISFACTION WITH LIVIN BY MANDIRI VIRTUAL CREDIT CARD FEATURES

2025-03-24T01:15:09+00:00

Credit cards are a payment method that replaces cash, enabling transactions at various locations that accept them. Credit cards can be physical or virtual, with virtual cards facilitating transactions without the need for a physical card. Bank Mandiri offers virtual credit cards to support digital transactions, providing benefits such as discounts and cashback. Although Bank Mandiri experienced a decline in credit card usage during the pandemic, there has been a post-pandemic upward trend influenced by the features of the virtual credit card in Livin' by Mandiri. This is attributed to customer satisfaction, which has become a key factor in this increase. However, as of now, there has been no research conducted on customer satisfaction regarding the virtual credit card feature in Livin' by Mandiri. Therefore, this study aims to analyze customer satisfaction levels with the virtual credit card feature using the PIECES Method (Performance, Information, Economy, Control, Efficiency, and Service). To support the theoretical framework of this issue, relevant journal reviews and data collection through questionnaires were conducted, with the population size based on Bank Mandiri's annual reports for 2022–2023. The analysis of the questionnaire data, employing the PIECES method and the Likert Scale, revealed the highest scores in the categories of Performance and Service, both scoring 4.20. The Efficiency category scored 4.19, while the Information and Control categories each scored 4.15. The lowest score was observed in the Economy category, with a score of 4.13. Therefore, it can be concluded that customers are satisfied.

ASSESSING SMPIT AJIMUTU GLOBAL INSANI WEBSITE QUALITY USING THE WEBQUAL 4.0 METHOD

2025-09-08T02:42:38+00:00

Digitalization in the world of education encourages schools to have quality websites to provide online information and learning services. This study aims to measure the quality of the SMPIT Ajimutu Global Insani website using the Webqual 4.0 method, which involves three main dimensions: usability quality, information quality, and service interaction quality. This research method involves a survey of 50 respondents consisting of teachers, students, and parents of students. Data were analyzed descriptively using a Likert scale to evaluate the level of user satisfaction. The results showed that the information quality dimension had the highest score (4.2), followed by service interaction quality (4.0), while usability quality scored the lowest (3.8). These findings indicate that the website content is relevant, but navigation and interface design need improvement. Recommendations are given to improve the quality of the website, including optimizing interactive features and adding multimedia content. The implementation of the results of this study is expected to support the digital transformation of schools more effectively.

EVALUATION OF USER PERCEPTIONS AND SATISFACTION THROUGH SENTIMENT ANALYSIS NEWS APPLICATIONS WITH NAIVE BAYES

2025-09-23T09:18:54+00:00

The development of digital technology has driven the transformation of mass media into online news platforms such as Detikcom, Kompas.id, and CNN Indonesia. Competition among these news applications has created the need to evaluate user perceptions of service quality. This study aims to analyze user sentiment toward the three news applications based on reviews from the Google Play Store. The methods employed include web scraping, text pre-processing, labeling using the IndoBERT model, feature extraction with the TF-IDF method, and sentiment classification with the Naive Bayes algorithm. To address class imbalance in the dataset, the Synthetic Minority Over-sampling Technique (SMOTE) was applied. Model evaluation was conducted using accuracy, precision, recall, and F1-score metrics. The results show that the Naive Bayes model achieved high accuracy, namely 88.5% for Kompas.id, 88.8% for Detikcom, and 90.8% for CNN Indonesia. The analysis also revealed that positive reviews are more dominant, although recurring criticisms were identified regarding advertisements and technical performance of the applications. The use of Generative AI further assisted in automatically summarizing opinions and sentiment patterns. These findings provide valuable insights for developers in enhancing user experience and refining the features of digital news applications

OPTIMIZING PHARMACEUTICAL DISTRIBUTION IN PUBLIC HEALTH CENTERS USING FUZZY C-MEANS CLUSTERING

2025-09-19T09:31:04+00:00

Efficient drug distribution is fundamental to ensuring the quality of public healthcare services. However, health departments often face challenges with imbalances between drug demand and available supply. This study addresses this issue by applying the Fuzzy C-Means (FCM) clustering algorithm to categorize drug demand levels across 16 public health centers (puskesmas) in Langkat Regency, Indonesia, from 2021 to 2023. Using historical data from 2,400 drug records, the analysis identified five distinct demand clusters: Very Low, Low, Medium, High, and Very High. The results revealed a significant disparity in drug needs, with the "Very High" demand cluster dominating (51.29% of data) in centers like Besitang and Tanjung Selamat, driven by high morbidity rates. In contrast, other clusters were less prevalent, such as the "Low" demand cluster, which was primarily concentrated in the Gebang health center. These findings, visualized using t-SNE plots, highlight significant regional variations in pharmaceutical needs. This data-driven clustering provides a robust framework for the Langkat District Health Office to develop more targeted, efficient, and equitable drug distribution strategies, ultimately improving healthcare service delivery.

OPTIMIZED DEEP AUTOENCODER WITH L1 REGULARIZATION AND DROPOUT FOR ANOMALY DETECTION IN 6G NETWORK SLICING

2025-10-01T04:21:31+00:00

The increasing complexity of 6G network slicing introduces new challenges in identifying abnormal behavior within highly virtualized and dynamic network infrastructures. This study aims to address the anomaly detection problem in 6G slicing environments by comparing the performance of three models: a supervised random forest classifier, a basic unsupervised autoencoder, and an optimized deep autoencoder enhanced with L1 regularization and dropout techniques. The optimized autoencoder is trained to reconstruct normal data patterns, with anomaly detection performed using a threshold- based reconstruction error approach. Reconstruction errors are evaluated across different percentile thresholds to determine the optimal boundary for classifying abnormal behavior. All models are tested on a publicly available 6G Network Slicing Security dataset. Results show that the optimized autoencoder outperforms both the baseline autoencoder and the random forest in terms of anomaly sensitivity. Specifically, the optimized model achieves an F1- score of 0.1782, a recall of 0.2095, and an accuracy of 0.714. These results indicate that introducing regularization and dropout significantly improves the ability of autoencoders to generalize and isolate anomalies, even in highly imbalanced datasets. This approach provides a lightweight and effective solution for unsupervised anomaly detection in next- generation network environments.

ANALYSIS OF THE CANTEEN INFORMATION SYSTEM AT AN-NAWAWI ISLAMIC BOARDING SCHOOL USING PIECES

2025-10-01T09:10:29+00:00

The canteen at An-Nawawi Modern Islamic Boarding School Bogor plays a crucial role in supporting students' daily activities; however, its current information system is not optimally integrated to handle the growing student population and operational complexity. This study aims to analyze the problems within the canteen's information system to identify priority areas for improvement. The methodology employs a quantitative analysis using the PIECES framework, which evaluates six variables: Performance, Information, Economy, Control, Efficiency, and Service. Data was collected by distributing a questionnaire to respondents via Google Forms. The research findings show that among the six variables, the Control variable obtained the lowest average score (563.4), indicating a significant weakness in the supervision and data security aspects of the current system. Therefore, it is concluded that the main priority for future development should be focused on strengthening the control aspects to create a more secure, reliable, and well-managed canteen information system.