A Reputation Synthesis Method Based on Multiple Evaluation Criteria

Abstract

The invention discloses a multi-evaluation criterion-based credit synthetic method, which comprises the following steps of a, setting a grading section of a grading system according to need; b, setting an upper limit N for the number of recent grading records for other nodes needing to be saved by each node; c, selecting N recent grading records for other nodes of a node x, and computing the trust propensity of the node x; d, when the node x receives an interacting request from another node y, arbitrarily selecting at most M candidate nodes which have the same level of trust propensity as the node x and can provide grading for the node y from recommended nodes by the node x, and acquiring the latest grading for the node y by the nodes; e, performing grading computing on the grading obtained in step d, and finally computing credibility of the node y. By the method, the influence of different criteria on a model under the same recognition framework is avoided.

Description

technical field [0001] The invention relates to the technical field of Internet and information security, in particular to a reputation synthesis method based on multiple evaluation standards. Background technique [0002] Currently, reputation models are mainly used in the Internet, peer-to-peer (Peer-to-Peer), self-organizing networks, etc. Typical calculation methods include simple accumulation or averaging. The cumulative trust model takes the sum of all the scores obtained by a node as the overall trust degree of the node. This is the most basic way of merging trust evidence, but it is difficult to reflect the true credibility of the evaluated node. When a node gets two diametrically opposed ratings, direct accumulation will cause the two ratings to cancel each other out, resulting in a medium trust score. The average trust model takes the average score obtained by a node as the credibility of the node, but the defect of this model is that the ability to distinguish t...

AI Technical Summary

Problems solved by technology

The average trust model takes the average score obtained by a node as the credibility of the node, but the defect of this model is that the ability to distinguish the trust status is insufficient. The obtained score is in an intermediate state, and other nodes cannot obtain valid information from such a score

[0003] The above processing methods did not give enough consideration to the relevant factors that affect trust, and only purely mathematically superimposed the evidence. Since there is no standard for measuring multiple evaluations, the scores given do not represent the actual reputation of the evaluators.

Images