Community credibility evaluation method based on improved D-S evidence theory

Abstract

The invention discloses a community credibility evaluation method based on an improved D-S evidence theory. The method comprises the steps of calculating a trust function and a likelihood function ofeach community by utilizing evaluation of basic probability allocation (BPA) on an evidence source; secondly, constructing an interval number through the trust function and the likelihood function toobtain a trust interval of each community; then, obtaining descending sort of the interval number of the communities through aggregation of an ordered weighted average operator; and finally, sorting the aggregated results to obtain a credibility evaluation result of each community. According to the method, the overall credibility of the community can be accurately assessed, and the credibility ofthe community can still be efficiently and accurately assessed even under the conditions that the data types or characteristics of the nodes are unknown, the nodes are attacked to different degrees ordata are forged; according to the method, the nodes in the network are fused according to the evidences to obtain the support degree of all the evidences for each community, so that the credibility of the communities can be effectively evaluated.

Description

technical field [0001] The invention belongs to the field of network security and credibility assessment, in particular to a community credibility assessment method based on improved D-S evidence theory. Background technique [0002] In recent years, with the rapid development and wide application of computer and network technology, the degree of informatization of society has increased, making computer network information storage, communication and processing an indispensable part of people's life and work. With the development of my country's Internet, the number of Internet users has exploded, and the requirements for network security and credibility have become more and more stringent. In fact, real social networks are typical complex networks. Therefore, we can analyze the credibility of the network according to the method of complex network analysis. Since community structure is a main feature of complex networks, we can start with the analysis of community credibili...

AI Technical Summary

Problems solved by technology

[0005] Chinese patent CN201910204538.7 discloses a method for evaluating the credibility of complex simulation systems based on network topology paths, specifically: first, analyze the information interaction relationship between component models in complex simulation systems, and calculate the weight of edges between nodes based on objective indicators , abstract the complex simulation system into a directed and authoritative model interaction network; calculate the out-degree of each node in the model interaction network, select the node with a larger out-degree as the initial node, and start from the initial node to obtain different single execution path, and calculate the credibility of the single execution path, the credibility of the simulation system can be obtained by integrating the credibility of all execution paths; however, this method is to quantify the entire The reliability of the complex simulation system; therefore, the evaluation method of the simulation system is limited by the reliability of the single-point model node, which needs to meet the precondition that the reliability of the single-model node is known. The calculation is complex and the process is cumbersome, and the accuracy of the final result depends on the accuracy of the initially selected model, so the practicability of this method is not high;

[0006] Chinese patent CN200910079036.2 discloses a node credibility evaluation algorithm based on Kirchhoff's theorem; it includes the following steps: first, among multiple nodes in the network, any two nodes with a trust relationship are connected to form Resistor network; then apply Kirchhoff's theorem to calculate the resistance value of each node, that is, the trust degree of the node; then recalculate the resistance value of each node according to the calculated resistance value of each node, and this process is repeated many times , and perform normalization until it converges to the set threshold, so as to obtain the credibility of the community; however, there are the following problems in this scheme: First, the choice of trust relationship has an important impact on the detection results, and choosing different trust relationships It may make the detection results different; the second is to connect the nodes with trust relationship, and then repeatedly calculate the credibility of each node, the amount of calculation is huge, and the implementation is more complicated;

[0007] At present, most of the existing community credibility assessment methods determine the credibility of the entire community by dividing nodes with trust relationships and exploring the credibility of each node in the node group; these methods rely on established rules or reasoning model, and in the face of the following situations, the credibility of the community cannot be calculated well: (1) the type or characteristics of the nodes are unknown, and there are no nodes with relevant trust relationships; (2) the nodes may be attacked to varying degrees Or data falsification and other behaviors, resulting in credibility evaluation errors

Images