Internet time-delay prediction based on Wavelet transformation and ARIMA

Abstract

Based on the analysis of the fundamental characteristics of the end-to-end network time delay, this paper presents a new network delay prediction method based on wavelet- autoregressive integrated moving average model (W-ARIMA). The wavelet filter is applied to eliminate the noise signal from the time delay data and then the ARIMA model is used to predict the time delay sequence. An experimental example is given to show that the performance of the proposed delay predicting method, the results show that the W-ARIMA proposed in this paper has much higher prediction accuracy compared with the method based on the direct autoregressive integrated moving average (ARIMA) model.