2024
1.
Al-Thaedan, Abbas; Shakir, Zaenab; Mjhool, Ahmed Yaseen; Alsabah, Ruaa; Al-Sabbagh, Ali; Nembhard, Fitzroy; Salah, Monera
A machine learning framework for predicting downlink throughput in 4G-LTE/5G cellular networks Journal Article
In: International Journal of Information Technology, vol. 16, pp. 651–657, 2024, ISSN: 2511-2112.
Abstract | Links | BibTeX | Tags: cellular networks, machine learning
@article{nokey,
title = {A machine learning framework for predicting downlink throughput in 4G-LTE/5G cellular networks},
author = {Abbas Al-Thaedan and Zaenab Shakir and Ahmed Yaseen Mjhool and Ruaa Alsabah and Ali Al-Sabbagh and Fitzroy Nembhard and Monera Salah },
editor = {M. N. Hoda},
doi = {https://doi.org/10.1007/s41870-023-01678-w},
issn = {2511-2112},
year = {2024},
date = {2024-01-02},
urldate = {2024-01-02},
journal = {International Journal of Information Technology},
volume = {16},
pages = {651–657},
abstract = {The current and next generations of cellular networks produce a massive amount of data. With this vast parameter increase, cellular communication networks have grown incredibly complicated. In addition, these cellular networks are unmanaged with conventional techniques, and a more advanced design and optimization methodology that depends on Machine Learning (ML) models is necessary. This work proposes a framework model for predicting downlink throughput (DL-Throughput) using ML models in fourth and fifth generations (4G/5G) cellular networks. The important parameters are selected from data measurements based on the correlation coefficient. The critical and effective parameters such as Reference Signal Received Power (RSRP), Signal to Interference and Noise Ratio(SINR), Received Signal Strength Indicator (RSSI), and Reference Signal Receive Quality (RSRQ) have been applied for the training model to predict the DL-Throughput in cellular networks. The prediction accuracy of the determination coefficient ranges between 89% and 96% from three different operators.},
keywords = {cellular networks, machine learning},
pubstate = {published},
tppubtype = {article}
}
The current and next generations of cellular networks produce a massive amount of data. With this vast parameter increase, cellular communication networks have grown incredibly complicated. In addition, these cellular networks are unmanaged with conventional techniques, and a more advanced design and optimization methodology that depends on Machine Learning (ML) models is necessary. This work proposes a framework model for predicting downlink throughput (DL-Throughput) using ML models in fourth and fifth generations (4G/5G) cellular networks. The important parameters are selected from data measurements based on the correlation coefficient. The critical and effective parameters such as Reference Signal Received Power (RSRP), Signal to Interference and Noise Ratio(SINR), Received Signal Strength Indicator (RSSI), and Reference Signal Receive Quality (RSRQ) have been applied for the training model to predict the DL-Throughput in cellular networks. The prediction accuracy of the determination coefficient ranges between 89% and 96% from three different operators.