The Evolution Mechanism And Application Of Hypernetwork

The emergence of complex networks makes people's cognition of the world more intuitive and provides new ideas for the research of complex systems and complex world.However,with the development of the world and the arrival of the era of big data,the world is getting more and more complex,and the network is getting more complex,the network scale is larger,the number of nodes and connecting sides is more,the types of nodes in the same network are also different,using complex network theory can not solve these complex network problems.The emergence of the hypernetwork just solves this problem.The underlying topology of the hypernetwork is a hypergraph.A hyperedge of a hypergraph can contain two or more nodes.A hyperedge represents a multi-dimensional relationship between nodes.You can express more information in the network.In this thesis,the main research object is the hypergraph-based hypernetwork,based on the construction of the hypernetwork model and the hypergraph theory,the priority connection problem in the hypernetwork model construction is studied,and then the practical application of the hypernetwork model construction is discussed and analyzed.The main contents are as follows:1.The priority connection method in the construction of hypernetwork model is studied.This thesis researches how to realize the optimal selection in the evolution process of the hypernetwork,and proposes two methods: the roulette wheel method and the linked list method to achieve priority connection.Through simulation experiments,some characteristics of the uniform hypernetwork and the non-uniform hypernetwork are analyzed.The time of the priority connection process between the roulette method and the linked list method is compared.2.In view of some problems existing in the evolution mechanism of the current hypernetwork model,the evolution mechanism in the process of building the existing hypernetwork model is improved,and a hypernetwork model based on logistic regression is constructed.Three main improvements have been made:(1)adding the maximum capacity of the hyperedge of the hypernetwork;(2)replacing thepreferential choice to handle the connection between nodes with a logistic regression function;(3)the number of connections of newly added nodes has been increased in a sub-linear manner functions instead of constants.Then,some laws of hyper-degree power-law distribution of the hypernetwork based on logistic regression are analyzed through simulation experiments.3.In order to study the relationship between key words and related scholars,this thesis constructs a keywords-scholars hypernetwork model,and measures some indicators based on the hypernetwork model,summarizes some characteristics of the keywords-scholars hypernetwork model.