Comparison between artificial neural network and Non-linear Auto Regressive Moving Average (NARMA) models for internet traffic / Muhammad Naqiuddin Sahrani

• Main Author: Sahrani, Muhammad Naqiuddin
• Format: Article
• Language: English
• Published: 2015
• Subjects: Neural networks (Computer science); Computer network protocols
• Online Access: https://ir.uitm.edu.my/id/eprint/40582/1/40582.pdf; https://ir.uitm.edu.my/id/eprint/40582/
• Institution: Universiti Teknologi Mara

This paper compare the performance of Artificial Neural Network with Non-Linear Auto Regressive Moving Average (NARMA) models by predicting internet traffic The ANN models that used in this paper was Non-Linear Auto Regressive (NAR) model where this model was train to predict the internet traffic The NARMA model result was obtained by combine the result of NAR and Moving Average (MA) models where the result of MA model obtain by predicting the residual of NAR model. The performance of both models was evaluated by using model fitting and model validation tests. NARMA models show better results in internet forecasting.