Predicting mobile networks' behaviors using Linear Neural Networks
Cuong Pham‐Quoc, Tuan Tang-Anh · 2021
Predicting cellular networks' behaviors for making network development strategies such as 3G/4G/5G technology layers, radio resource management and optimizing Capital Expenditure (CAPEX) and Operating Expenditure (OPEX) is a very important task. This paper proposes using a Linear Neural Network which has an acceptable accuracy for predicting networks' behaviors in cellular systems based on real statistical information from Operations Support Systems (OSS). Training and testing data are taken from real 4G/LTE networks for verifying and analyzing the results. Once the prediction is done then corresponding configuration tunings are carried in the network to fit service requirements and leverage network's performance.