Research On Several Problems Of Node Ranking In Complex Networks

The rise and development of the Internet has laid the foundation for the emergence of complex network systems.As an abstract representation of a real network system,the complex network has become an important tool for researching network systems because of its logical relationship and flexible structure changes.In different network systems,the sequencing problem aimed at each unit concerned is of great significance to system administrators.For example,the abnormal degree sequencing of network nodes could effectively detect mistakes so as to help administrators carry out fault detection and diagnosis.And the ranking of user influence in social networks can help managers understand the network information dissemination and maximize spread.In this paper,based on the actual research work,the author studied several problems of node ranking in two representative complex network systems:The first thing to study is the distributed network system.The information in this system is the data flow generated by the hardware.This study models the data flow to form a complex network,this paper also studied the problems related to the ordering of abnormality of the nodes at the same time.The specific steps are as follows:Based on the ARX model,an efficient edge search algorithm is proposed.The algorithm uses a segmentation regression model.On the basis,this study uses the partitioned regression matrix to calculate the standard fitness score.The standard fitness score continuously reduces the candidate invariant edges of each segmented portion,thereby achieving the effect of improving the temporal performance without losing the invariant edge.Another representative complex network system is the social network.The information in this type of system is generated by the user.This dissertation has conducted research on issues concerning the level of user activity in this type of network.The specific work content is as follows:In a social network system that contains a large amount of data,the social network data is modeled using misdirected graphs by analyzing the characteristics of the social network,and the user's initial vitality score is designed based on the spatial and temporal correlation of nodes.The algorithm,afterwards,this study processed the original vitality score through an iterative algorithm,making the user's vitality score more stable,and using the user vitality score to rank the nodes in the social network.According to the sorting result,taking the actual application requirements as the starting point,the exponential smoothing forecasting algorithm is adopted.Based on this,the algorithm is improved according to the characteristics of our sorting algorithm,and a simple and effective user vitality prediction algorithm is formed.I hope this research can provide some optimization ideas for the workers in the Internet field and can be practically applied in actual production,at the same time,I also hope it can create theoretical conditions for the further development of the Internet industry of China.