Small-time scale network traffic prediction based on a local support vector machine regression model

In this paper we apply the nonlinear time series analysis method tosmall-time scale traffic measurement data. The prediction-basedmethod is used to determine the embedding dimension of the trafficdata. Based on the reconstructed phase space, the local supportvector machine prediction method is used to predict the trafficmeasurement data, and the BIC-based neighbouring point selectionmethod is used to choose the number of the nearest neighbouringpoints for the local support vector machine regression model. Theexperimental results show that the local support vector machineprediction method whose neighbouring points are optimized caneffectively predict the small-time scale traffic measurement dataand can reproduce the statistical features of real trafficmeasurements.