LSTM Neural Network Architecture and Hyperparameter Exploration for Handover Simulation in 5G Network

EasyChair Preprint 11653

Abstract

This paper presents a machine learning model for optimizing a handover process in 5G networks. The data for learning and testing is simulated using NS3. By using RNN with LSTM layer, model is enable to decide which cell to handover to provide the highest download success rate. Key of this analysis is exploring the hyperparameter of the model such as hidden nodes, epoch, dropout rate to provide the highest download success rate.

Keyphrases: Dropout, Epoch, Handover, LSTM, RNN, hidden nodes

@booklet{EasyChair:11653,
  author    = {Ali Zayn Murteza and W Abdullah Rafa and Iskandar and Baud Prananto},
  title     = {LSTM Neural Network Architecture and Hyperparameter Exploration for Handover Simulation in 5G Network},
  howpublished = {EasyChair Preprint 11653},
  year      = {EasyChair, 2024}}