Research On Source Tracing Algorithm Based On Node Importance

Node importance and information traceability have always been important research contents in the field of social networks.Due to the poor performance of accuracy and discrimination in measuring node importance only from a single attribute,many scholars focus on multi-attribute fusion for measurement.However,most methods ignore the problem that different attributes contribute differently to the importance of nodes,resulting in limited performance improvement;in terms of information traceability,the existing traceability algorithms ignore the structural characteristics of the network itself and the influence of prior estimation on the traceability results,resulting in low accuracy and large error distance of the traceability results.In view of the above problems,the following two aspects are studied in this paper:Firstly,a ranking algorithm of node importance based on node location information and structural attributes is proposed.On the one hand,the algorithm improves the computing performance of node position attributes by using the iterative information generated in each iteration process of K-Shell decomposition;on the other hand,the network structure entropy is used to measure the structural attributes of the nodes,and the importance influence of the neighbor nodes on the nodes is evaluated according to the extended neighborhood,which further improves the discriminating ability of the algorithm.Finally,the entropy weight method is used to calculate the different weights of the location information and the structure attribute,which makes the sorting result more reasonable.Experimental results show that the algorithm has good performance in terms of accuracy,discrimination and computational efficiency.Secondly,combining node importance with information traceability,a traceability algorithm combining a prior estimate and a posterior estimate is proposed.On the one hand,the algorithm sorted the importance of infected nodes in the infected network,and processed the ranking value,so that it could overcome the influence of the nodes with higher intermediate value in the original network on the infected network,so as to obtain the prior estimate of the traceability algorithm;on the other hand,the propagation centrality algorithm is used as a posterior estimate to obtain the information traceability algorithm which fuses the prior estimate and the posterior estimate.Experimental results show that the proposed traceability algorithm combining prior and posterior estimates has the advantages of higher accuracy and smaller error distance.Finally,on the basis of the above theoretical research,taking the real data set in social network as an example,the prototype system of information traceability is designed and implemented.Adopts the modularization development method,completes the design of the business process and the system page,realizes the visualization function of the algorithm.In this thesis,there are 19 figures,8 tables and 92 references.