Traffic Forecasting of Core Network Based on Improved Logistic Regression

Abstract

Traffic forecasting of core network plays an important role in network planning, traffic management, etc. Therefore, a predictive model that can accurately predict core network traffic is needed. This article proposes a new traffic forecasting method of core network based on logistic regression (LR). In order to get an accurate logistic model, new LR parameter estimation algorithm is proposed. First, the unknown parameters of LR are replaced by the minimum variance unbiased estimator to ensure the accuracy. In order to reduce the computational complexity, a statistical model of LR is introduced. Then, the unknown parameters of the LR are estimated based on the Cramer-Rao lower bound, and then the LR is further obtained based on the proposed estimator. Finally, the accuracy of the model is verified through experiments based on traffic data of core network. Experimental result shows that the improved logistic model proposed in this paper is more accurate than other methods.