Development of Internet Traffic Prediction Software Using Time-Series Multilayer Perceptron
Abstract
Internet traffic prediction plays a fundamental role in network design, management, control and optimization. Although there exist several studies in litereture that focus on predicting Internet traffic using statistical and machine learning methods, to the best of our knowledge, a fully functional off-the-shelf software with different optimization capabilities has not been developed. The purpose of this study is to develop a new software for prediction of Internet traffic data based on time-series Multilayer Perceptron (MLP). The software includes features such as the optimization of the number of hidden layers and neurons in each layer and feedback delay optimization with respect to autocorrelations. The Internet traffic data from two different Internet Service Providers, varying by 1-hour and 5-minute time frequencies, have been used for testing the software. The datasets have been split into training and testing sets via 70-30% and 80-20% split ratios. The Mean Absolute Percentage Error (MAPE) has been utilized as the main error rate metric in order to evaluate the accuracy of the prediction models. It has been observed that the MAPE's of the Internet traffic prediction models change between 3.25 and 9.09. One can conclude that the developed software can be used for Internet traffic prediction within acceptable error rates.
Keywords
Kaynakça
- Bian, G., Liu, J., ve Lin, W. 2017. Internet Traffic Forecasting using Boosting LSTM Method. DEStech Transactions on Computer Science and Engineering.
- Chabaa, S., Zeroual, A. ve Antari, J., 2010. Identification and Prediction of Internet Traffic Using Artificial Neural Networks. Journal of Intelligent Learning Systems and Applications. 2(3), 147-155.
- Cortez, P., Rio, M., Rocha, M. ve Sousa, P., 2012. Multi-scale Internet Traffic Forecasting Using Neural Networks and Time Series Methods. Expert Systems. 29(2), 143-155.
- Daly, C. 2016. H. 265 Video Traffic Prediction Using Neural Networks. Georgia Southern University Research Symposium. s127.
- Jiang, M., Wu, C. M., Zhang, M. ve Hu, D. M., 2009. Research on the Comparison of Time Series Models for Network Traffic Prediction. Acta Electronic Sinica. 37(1), 2353-2358.
- Katris, C., ve Daskalaki, S. 2015. Comparing forecasting approaches for Internet traffic. Expert Systems with Applications, 42(21), 8172-8183.
- Liu, X., Fang, X., Qin, Z., Ye, C., ve Xie, M. 2011. A Short-Term Forecasting Algorithm for Network Traffic Based on Chaos Theory and SVM. Journal of Network and Systems Management. 19(4), 427-447.
- Narejo, S., ve Pasero, E. 2018. An Application of Internet Traffic Prediction with Deep Neural Network. In Multidisciplinary Approaches to Neural Computing. pp. 139-149. Springer, Cham.
- Oliveira, T. P., Barbar, J. S. ve Soares, A. S. 2014. Multilayer Perceptron and Stacked Autoencoder for Internet Traffic Prediction. In Network and Parallel Computing. pp. 61-71.
- Sahrani, M. N., Zan, M. M. M., Yassin, I. M., Zabidi, A., ve Ali, M. S. A. M. 2017. Artificial Neural Network Non-linear Auto Regressive Moving Average (NARMA) Model for Internet Traffic Prediction. Journal of Telecommunication, Electronic and Computer Engineering. 9(1-3), 145- 149.