Abstract:
|
In order to improve the accuracy of the multiple influence matrix evaluation method for the identification of important nodes in the directed weighted network, the existing influence matrix is not considered node efficiency, so optimizing the influence matrix construction method.There is a certain subjectivity in the process of using traditional weighting method to weight the influence matrix to form multiple influence matrix. Based on this, a weight self-learning algorithm is proposed to obtain the weight of influence matrix. This method can traverse all weight combinations in the weight value space and determine the optimal combination according to the influence of different combinations on the node importance value, eliminating the subjective factors in the process of weighting. The method is applied to simulation network and ARPA network. The results show that the method can accurately identify the important nodes in the network. Finally, the experimental results are compared with the simulation results of network node connectivity to further verify the effectiveness of this method.
|