Garuda - Garba Rujukan Digital

Implementation of the Support Vector Machine (SVM) Algorithm to Improve the Accuracy of Computer Network Performance Predictions Desi Irfan; Fahruzi Sirait; Rahadatul, Aisy Riadi; Aldi Indrawan; Juni Purwanto
International Journal of Health Engineering and Technology Vol. 4 No. 1 (2025): IJHET May 2025
Publisher : CV. AFDIFAL MAJU BERKAH

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55227/ijhet.v4i1.271

Computer network performance is very important in supporting various digital activities, but systems often cannot accurately predict changes in performance, which can cause service disruptions and economic losses. This research aims to implement the Support Vector Machine (SVM) algorithm to increase the accuracy of network performance predictions based on parameters such as latency, packet loss, throughput and jitter. Data is collected through network simulation and real data monitoring, then processed with normalization and selection of relevant features. The SVM model is tested with various kernels, including linear, RBF, and polynomial, to find the best configuration. Performance evaluation uses accuracy, precision, recall, F1-score, and ROC-AUC metrics, with cross-validation to increase the reliability of the results. The results show that the RBF kernel provides a prediction accuracy of 92%, higher than baseline methods such as Decision Tree and Logistic Regression. This model shows its potential to be applied in computer network monitoring systems to predict network performance in real-time, with the possibility of wider implementation in artificial intelligence-based network applications. Therefore, this research not only contributes to machine learning theory in the field of computer networks, but also provides practical solutions that can improve the management and optimization of network performance in various environments that require fast and accurate data processing

Analysis of risk factors for failure of hypertension therapy based on medical history and drug consumption using Random Forest Desi Irfan; Novica Jolyarni; Halimah Tusakdiyah Harahap; Baginda Restu Al Ghazali; Riswan Syahputra Damanik
International Journal of Health Engineering and Technology Vol. 2 No. 4 (2023): IJHET NOVEMBER 2023
Publisher : CV. AFDIFAL MAJU BERKAH

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55227/ijhet.v4i1.276

Computer network performance is very important in supporting various digital activities, but systems often cannot accurately predict changes in performance, which can cause service disruptions and economic losses. This research aims to implement the Support Vector Machine (SVM) algorithm to increase the accuracy of network performance predictions based on parameters such as latency, packet loss, throughput and jitter. Data is collected through network simulation and real data monitoring, then processed with normalization and selection of relevant features. The SVM model is tested with various kernels, including linear, RBF, and polynomial, to find the best configuration. Performance evaluation uses accuracy, precision, recall, F1-score, and ROC-AUC metrics, with cross-validation to increase the reliability of the results. The results show that the RBF kernel provides a prediction accuracy of 92%, higher than baseline methods such as Decision Tree and Logistic Regression. This model shows its potential to be applied in computer network monitoring systems to predict network performance in real-time, with the possibility of wider implementation in artificial intelligence-based network applications. Therefore, this research not only contributes to machine learning theory in the field of computer networks, but also provides practical solutions that can improve the management and optimization of network performance in various environments that require fast and accurate data processing.

Simulation and Detection of Phishing Attacks on Student Academic Emails Using Social Engineering Techniques Santosa Pohan; Desi Irfan; Intan Nur Fitriyani; Yusril Iza Mahendra Hasibuan; Indah Chayani
International Journal of Health Engineering and Technology Vol. 2 No. 4 (2023): IJHET NOVEMBER 2023
Publisher : CV. AFDIFAL MAJU BERKAH

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55227/ijhet.v2i4.283

Phishing attacks on student academic emails are a serious threat to information security. Social engineering techniques are often used in these attacks to manipulate victims into divulging sensitive information, such as passwords and other personal data. This research aims to analyze and detect phishing attacks that use social engineering techniques on student academic emails. In this research, a phishing attack simulation was carried out with the scenario of falsifying the identity of an academic institution and creating fake emails that appear legitimate. Students as simulated subjects were tested to see how they reacted to deceptive phishing emails, such as clicking on malicious links or downloading infectious attachments. The detection methods used include heuristic analysis and machine learning techniques, where the system is trained to recognize suspicious patterns in emails, including elements such as unusual subjects, links and attachments. The research results show that phishing attacks that utilize social engineering are effective in manipulating victims. On the other hand, detection using machine learning and heuristic analysis can achieve a high level of accuracy in identifying phishing attacks. This research also underscores the importance of increasing awareness about cyber security among students as well as the need to develop more effective phishing detection tools.

Analysis of risk factors for failure of hypertension therapy based on medical history and drug consumption using Random Forest Desi Irfan; Novica Jolyarni D; Halimah Tusakdiyah Harahap; Baginda Restu Al Ghazali; Riswan Syahputra Damanik
International Journal of Health Engineering and Technology Vol. 2 No. 4 (2023): IJHET NOVEMBER 2023
Publisher : CV. AFDIFAL MAJU BERKAH

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55227/ijhet.v2i4.284

Cardiovascular disease is a major cause of global morbidity and mortality, with many patients experiencing therapy failure despite treatment. This study analyzes risk factors for failure of antihypertensive therapy based on medical history and drug consumption patterns using the Random Forest algorithm. Retrospective analytical research design using medical record data and structured interviews in hypertensive patients who have undergone treatment for at least one year. The dependent variable was therapy failure, defined as BP ≥140/90 mmHg despite treatment. Independent variables include medical history, drug consumption patterns, and demographic factors. Data is processed by handling missing data, normalization, and feature encoding. The Random Forest model was optimized using GridSearchCV and evaluated based on accuracy, precision, recall and AUC-ROC. Feature importance analysis identifies main risk factors, such as medication adherence, diabetes, and duration of hypertension. The model achieved 86% accuracy (AUC: 0.89), better than logistic regression (accuracy: 78%). These results confirm the importance of patient compliance and comorbidities in hypertension management. This study demonstrates the effectiveness of Random Forest in identifying high-risk patients, with recommendations for prioritization of interventions on medication adherence.

Pelatihan Deteksi Risiko Hipertensi Dengan Analisis Data Riwayat Medis Berbasis Random Forest Untuk Tenaga Kesehatan Masyarakat Desi Irfan; Evri Ekadiansyah; Halimah Tusakdiyah Harahap; Novica Jolyarni Dornik; Yusril Iza Mahendra Hasibuan
Sevaka : Hasil Kegiatan Layanan Masyarakat Vol. 1 No. 4 (2023): November: Sevaka : Hasil Kegiatan Layanan Masyarakat
Publisher : STIKES Columbia Asia Medan

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.62027/sevaka.v1i4.527

Hypertension is one of the most prevalent non-communicable diseases and a major risk factor for heart disease, stroke, and kidney disorders. The high prevalence of hypertension cases in the community, particularly in the working area of Puskesmas Kota Rantau Prapat, highlights the urgent need for more effective early detection efforts to prevent severe complications in the future. However, the limited capacity of healthcare workers in utilizing data analysis technologies has resulted in hypertension risk detection being dominated by conventional methods, which are often less accurate and inefficient. To address this issue, this community service program was conducted through training on the application of the Random Forest algorithm to analyze patients’ medical history data in order to detect hypertension risks. The training method included an introduction to the fundamentals of machine learning, data pre-processing stages, implementation of the Random Forest algorithm, and interpretation of prediction results. The outcomes of the program demonstrated that healthcare workers were able to understand the use of data analysis technologies to support more accurate early detection of hypertension. Furthermore, the participants gained practical skills in utilizing medical datasets to produce predictions that can serve as a decision-support tool for preventive medical actions.Thus, this training contributed to enhancing the capacity of community healthcare workers in integrating machine learning-based technologies into preventive healthcare services. This program is expected to serve as an initial step toward developing more effective, efficient, and sustainable data-driven health systems.

Title

Found 5 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 5 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 5 Documents
Search