Network Traffic Big Data Prediction Model Based on Combinatorial Learning

Abstract

Network traffic is an unstable time series. Under uncertain conditions and the lack of sufficient data, traffic prediction is a complex problem. The grey markov model is a prediction model combining the classical grey theory and the state escape behavior of markov chains. In this paper, water conservancy information network data set provided by a company is used as experimental data for modeling. Considering the fast computing power of the statistical model and the non-stationary characteristics of the network flow time series, this paper first constructs the gray markov Verhulst model. By taking the difference between the predicted results and the real data as a new data flow, GRU neural network algorithm is introduced to construct the composite model. In this paper, three model algorithms proposed based on actual network traffic data, LSTM and GRU are respectively used for simulation experiments, and the prediction effects of the three models are analyzed and evaluated. Experimental results show that the three short-term network traffic prediction models proposed in this paper can achieve good prediction results, among which the grey verhulst-markov GRU neural network model has the best prediction effect.