Claim Missing Document
Check
Articles

Found 2 Documents
Search
Journal : Journal of Applied Science, Engineering, Technology, and Education

Implementation of Machine Learning Algorithm with Extreme Gradient Boosting (XGBoost) Method In Hypertension Level Classification Rais, Zulkifli; Fahmuddin S, Muhammad; Saida, Saida; Triutomo, Agung
Journal of Applied Science, Engineering, Technology, and Education Vol. 7 No. 1 (2025)
Publisher : PT Mattawang Mediatama Solution

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35877/454RI.asci4191

Abstract

The increasing number of hypertension patients and the threat of serious complications make hypertension one of the leading causes of death worldwide. Early prevention is currently considered one of the best solutions. Early prevention through early detection can be achieved by utilizing machine learning technology. XGBoost is a machine learning algorithm based on gradient boosting machines. XGBoost applies regularization techniques to reduce overfitting and has faster execution speed as well as better performance. The objective of this research is to classify hypertension levels using the XGBoost method and leveraging hyperparameter tuning for optimization. In this study, the hyperparameter optimization technique used is gridsearchCV. The evaluation results of the XGBoost classification method using the best combination of parameters show good performance, where the XGBoost model achieves an accuracy of 93.3%, Precision of 97%, Recall of 92%, F1-Score of 93%, and AUC value of 0.935. This implies that the classification of hypertension levels in patients at Pelamonia Makassar Hospital can be well or accurately classified using the XGBoost method.
Statistical Dashboards and Business Intelligence in Campus Information Systems: A Bibliometric Review of Implementation Trends Ahmar, Ansari Saleh; Triutomo, Agung
Journal of Applied Science, Engineering, Technology, and Education Vol. 7 No. 3 (2025)
Publisher : PT Mattawang Mediatama Solution

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35877/454RI.asci2969

Abstract

Campus academic information systems have become mission-critical infrastructure in higher education, yet a significant paradox persists. Many universities operate monolithic architectures that consolidate student data and administrative functions within unified platforms—offering inherent security advantages including centralized authentication, unified access control, and simplified vulnerability monitoring. However, scholarly discourse examining how these secure integrated systems can simultaneously achieve advanced business intelligence capabilities remains remarkably thin. This bibliometric study analyzes 749 publications from Scopus (2010-2025) to map the intellectual landscape of campus information systems research, with particular attention to security frameworks and statistical dashboard implementations. The methodology combines linear regression trend analysis (β = 2.54, p = 0.00135, R² = 0.5317), Bradford's Law, Lotka's Law, and k-means clustering (k = 9). Results reveal statistically significant publication growth (CAGR = 1.46%), accumulating 6,039 citations (mean = 8.06) across 1,945 authors from 86 countries. Indonesia dominates contributions (26.1%), followed by China (10.3%) and the United States (8.1%). Thematic analysis identifies nine research clusters, with security-focused studies employing PTES and OWASP methodologies achieving 83% intrusion detection accuracy, while governance evaluations using COBIT and ISO 27001 reveal system maturity gaps. Critically, fewer than 10% of publications address real-time analytics or decision support visualization within secure monolithic architectures. The collaboration index (3.03 authors/document) and degree of collaboration (83.2%) indicate robust interdisciplinary practices. Findings suggest that while security research has matured, significant gaps persist in integrating business intelligence dashboards with secure monolithic systems—highlighting urgent need for research bridging data protection frameworks with analytical capabilities.