Station flow prediction method and device

Abstract

The invention discloses a station flow prediction method and device. The method comprises the steps of obtaining log data, calculating a predicted value based on the log data according to a Kalman filtering algorithm and a support vector machine (SVM) in order to obtain effective data more conveniently, calculating a weight by comparing an actual flow value, and constructing a parallel combinationmodel, establishing a prediction model according to a flow sequence in log data, obtaining predicted site network flow values, and using test data to check model prediction precision. By comparing results of Kalman prediction, SVM prediction and parallel combination model prediction based on Kalman and SVM, it is proved that a combination algorithm is adopted to predict network traffic, and the defect of single prediction of a traditional time sequence model is overcome; and the flow prediction result of the parallel combination model is reliable and effective and is more suitable for predicting the station flow. The parallel combination model can be applied to prediction of network traffic in a high-performance computing environment.

Description

technical field [0001] The invention belongs to the field of network flow monitoring, and in particular relates to a site flow prediction method and device. Background technique [0002] With the vigorous development of the Internet, the increase of site business volume and the continuous accumulation of users, the network traffic of the website presents complex and changeable characteristics, and the validity of the site traffic prediction results is increasingly demanding. Due to the characteristics of management system log traffic and forecasting requirements, the common single forecasting method can no longer meet the reliability and effectiveness requirements. The forecasting results of site traffic data are unreliable and invalid, which will affect the operating status of the site and network maintenance. Contents of the invention [0003] The purpose of the present invention is to provide a site traffic forecasting method and device to overcome the deficiencies of t...

AI Technical Summary

Problems solved by technology

[0002] With the vigorous development of the Internet, the increase of site business volume and the continuous accumulation of users, the network traffic of the website presents complex and changeable characteristics, and the validity of the site traffic prediction results is increasingly demanding. Due to the characteristics of the management system log traffic and forecasting requirements, the common single forecasting method can no longer meet the reliability and effectiveness requirements. The forecasting results of site traffic data are unreliable and invalid, which will affect the operating status of the site and network maintenance.

Images