In this study, a new model is developed that performs handover prediction in fifth generation small cell networks with Long Short Term Memory (LSTM) based deep neural networks. Firstly, the data set to be used for training in handover prediction was created with simulation scenarios designed in Riverbed Modeler. Through these scenarios, three input (RSSI, SNR and Jitter) variables and one output variable (desired value) were obtained to be used in the data set of the neural network. This data set was trained with machine learning algorithms LSTM, SVM, Tree, and Linear Regression techniques. LSTM-based deep neural network was compared with other regression algorithms and was found to have higher performance. When the test results of the trained model for LSTM are examined; R2 0.94, MAE 0.3315, MSE 0.3670, and RMSE value 0.6058 was found. It was observed that LSTM-based deep neural networks show high performance in regression processes. As a result, study shows that handover decisions can be predicted in 5G small cell networks with the proposed regression model.
: December 15, 2020
|APA||Cicioğlu, M . (2021). LSTM tabanlı Derin Sinir Ağı ile Beşinci Nesil Küçük Hücre Ağlarında El Değiştirme Tahmini . Bilecik Şeyh Edebali Üniversitesi Fen Bilimleri Dergisi , 8 (1) , 90-99 . DOI: 10.35193/bseufbd.840927|