Design and Implementation of a Machine Learning-Based Network Optimization Recommendation System with Web Performance Evaluation
Buletin Pos dan Telekomunikasi · 12.12.2025 · 23 (2) · 76-91
DOI: 10.17933/bpostel.v23i2.423 ↗Kurz gefasst
Fachartikel in der Zeitschrift Buletin Pos dan Telekomunikasi, online erschienen am 12.12.2025. Verfasst von Ahmad Rifai und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-SA 4.0).
Abstract (Originalsprache)
Network optimization is an essential process to maintain the quality of cellular services. However, manual analysis of drivetest data to determine optimization recommendations is time-consuming and inefficient. This study aims to develop a machine learning-based network optimization recommendation system implemented in the form of a website to assist RF Engineers in analyzing drivetest data more efficiently. The system uses a Random Forest Classifier to recommend the type of network optimization, achieving an average accuracy, precision, recall, and F1-score of 95.5%, and a Random Forest Regressor to predict network performance parameters after optimization, with an average R² of 0.9618, MAE of 0.0178, MSE of 0.00078, and RMSE of 0.0286. The dataset used is obtained from drivetests consisting of longitude, latitude, RSRP, SINR, Downlink Throughput, and Uplink Throughput parameters. The website was developed using the Flask framework and tested using System Usability Scale (SUS), Google Lighthouse, and GTmetrix. SUS testing obtained an average score of 79.16, categorized as “Good,” indicating that the website is easy to use and understand. Google Lighthouse testing obtained a performance score of 82, indicating good and responsive loading performance. GTmetrix testing showed an average performance score of 90.5% and structure score of 90.25%, indicating a well-structured website with optimal loading performance across various global server locations. This system can assist RF Engineers in analyzing drivetest data and making network optimization decisions more quickly, practically, and efficiently.
Zitieren
APA
Rifai, A., Wulandari, A., & Hikmaturokhman, A. (2025). Design and Implementation of a Machine Learning-Based Network Optimization Recommendation System with Web Performance Evaluation. Buletin Pos dan Telekomunikasi, 23(2), 76-91. https://doi.org/10.17933/bpostel.v23i2.423
Vancouver
Rifai A, Wulandari A, Hikmaturokhman A. Design and Implementation of a Machine Learning-Based Network Optimization Recommendation System with Web Performance Evaluation. Buletin Pos dan Telekomunikasi. 2025;23(2):76-91. doi:10.17933/bpostel.v23i2.423
BibTeX
@article{Rifai2025,
author = {Rifai, Ahmad and Wulandari, Asri and Hikmaturokhman, Alfin},
title = {Design and Implementation of a Machine Learning-Based Network Optimization Recommendation System with Web Performance Evaluation},
journal = {Buletin Pos dan Telekomunikasi},
year = {2025},
volume = {23},
number = {2},
pages = {76--91},
doi = {10.17933/bpostel.v23i2.423}
}
EUVIMED