Network traffic data filling method, device and equipment and storage medium

Abstract

The invention discloses a network traffic data filling method and device, equipment and a storage medium. The method comprises the steps: carrying out the modeling of network traffic data into a three-dimensional original tensor, deeply mining the periodic features between the network traffic data, and reflecting the multi-dimensional features of the network traffic data; except regression and CPdecomposition are combined to construct a loss function, accurate recovery of data can be carried out in a targeted mode by selecting a set weight w, and accurate recovery of elephant flow data is achieved. Meanwhile, Execut regression can describe the central characteristics of the data and can describe the tail characteristics of the data, the full-view characteristics of the data are reflected,and the problem that local characteristics of all parts of the data cannot be described through a traditional method is solved; according to the method, the factor matrix is updated according to thenon-negative matrix factorization algorithm and Except regression, in the updating process, it is not needed to calculate an inverse matrix of the matrix like an ALS algorithm, it is also not needed to repeatedly balance an appropriate learning step length like an SGD algorithm, and the calculation complexity is greatly reduced.

Description

technical field [0001] The invention belongs to the fields of computer technology and network technology, and in particular relates to a network flow data filling method, device, equipment and storage medium. Background technique [0002] Traffic matrix is ​​usually used to record traffic data between source nodes and target nodes, and is applied in network engineering scenarios such as load balancing, anomaly detection, and protocol design. But for a network with a complex structure, it is not an easy task to construct a traffic matrix by obtaining network traffic data between source nodes and target nodes. On the one hand, traffic data needs to be obtained by deploying and measuring physical devices at nodes. A complex network structure will generate many network nodes, and it is obviously unrealistic to deploy physical devices at each node. Q.Zhao et al. clearly pointed out this in "Robust trafficmatrix estimation with imperfect information: Making use of multiple dataso...

AI Technical Summary

Problems solved by technology

[0027] Aiming at the deficiencies of the prior art, the present invention provides a network traffic data filling method, device, equipment and storage medium to overcome the traditional tensor decomposition algorithm based on symmetric least squares, which mainly reflects the centrality of the data and cannot describe the data Defects in the local characteristics of each part to achieve accurate recovery of elephant flow data and reduce computational complexity

Images