Publication Type : Conference Paper
Publisher : 2017 International Conference on Advances in Computing, Communications and Informatics (ICACCI)
Source : 2017 International Conference on Advances in Computing, Communications and Informatics (ICACCI) (2017)
Keywords : Biological neural networks, computer network management, Deep learning, deep learning approaches, gated recurrent unit, GÉANT backbone networks, identity recurrent unit, Internet, learning (artificial intelligence), Logic gates, long short term memory, LSTM, Network management, Network traffic matrix, network traffic prediction, optimal network parameters, planning tasks, prediction, Predictive models, proactive approach, recurrent neural nets, recurrent neural network, Recurrent neural networks, RNN methods, RNN networks, telecommunication computing, telecommunication network planning, Telecommunication traffic, Time series, time series data modeling, traffic matrix estimation, Training
Campus : Coimbatore
School : School of Engineering
Center : Computational Engineering and Networking
Department : Computer Science, Electronics and Communication
Year : 2017
Abstract : Network traffic prediction aims at predicting the subsequent network traffic by using the previous network traffic data. This can serve as a proactive approach for network management and planning tasks. The family of recurrent neural network (RNN) approaches is known for time series data modeling which aims to predict the future time series based on the past information with long time lags of unrevealed size. RNN contains different network architectures like simple RNN, long short term memory (LSTM), gated recurrent unit (GRU), identity recurrent unit (IRNN) which is capable to learn the temporal patterns and long range dependencies in large sequences of arbitrary length. To leverage the efficacy of RNN approaches towards traffic matrix estimation in large networks, we use various RNN networks. The performance of various RNN networks is evaluated on the real data from GÉANT backbone networks. To identify the optimal network parameters and network structure of RNN, various experiments are done. All experiments are run up to 200 epochs with learning rate in the range [0.01-0.5]. LSTM has performed well in comparison to the other RNN and classical methods. Moreover, the performance of various RNN methods is comparable to LSTM.
Cite this Research Publication : R. Vinayakumar, Dr. Soman K. P., and Poornachandran, P., “Applying Deep Learning Approaches for Network Traffic Prediction”, in 2017 International Conference on Advances in Computing, Communications and Informatics (ICACCI), 2017.