Direct Forecast Method Based on ANN in Network Traffic Prediction

摘要

In this chapter, ARIMA (autoregressive integrated moving average) model and direct and iterative forecast methods based on ANN (artificial neural network) are adopted to fit and forecast the network traffic sequences. Different methods for predictive modeling are adopted to deal with the actual network traffic flow at different time intervals. With the GRA (gray relational analysis) method, the comparison and analysis of performance of the model show that the prediction error will be less if we use direct method for predictive modeling.