The Spatio-Temporal Dependent Social Network Model And Its Applications

The traditional social network analysis is a research method in the field of sociology.It explores the characteristics of nodes and networks by studying relational data.However,traditional social network analysis focuses on the characterization of nodes and network characteristics,ignoring the impact of other continuous response variables of nodes.For network related data with continuous response variables,network vector autoregression is usually used for analysis.However,this method does not take into account the spatial distance of nodes,and in the processing of network relations,the network averages the impact of response variables of adjacent nodes.For spatiotemporal network relational data,this paper proposes a spatiotemporal dependent social network model.This model not only overcomes the shortcomings of the average influence of the response variables of adjacent nodes,increases the estimation of individual effects,but also takes into account the influence relationship of the spatial distance of nodes,combines network relational data with spatial geographic data,and flexibly depicts the temporal and spatial dependence of the response variables of nodes.This paper integrates social network analysis and spatial econometric model,and analyzes the network from two dimensions of time and space.Therefore,the main contributions of this paper can be summarized as follows:1.This paper solves the problem of averaging network effects,highlighting the impact of individual node effects.Network Vector Autoregression(NVAR)model takes into account the effects of specific node response variables,network effects,momentum effects and residuals.However,NVAR model only considers the average network impact of the response variables of adjacent nodes on the individual nodes,and cannot reflect the individual impact of specific adjacent nodes.Therefore,on the basis of NVAR model,this paper broadens the network relationship influence averaged by response variables of adjacent nodes into individual effect influence.2.The model proposed in this paper depicts the temporal dependence while fully taking into account the influence of the spatial geographical structure of nodes,and depicts the spatial dependence.Although the Network Vector Autoregression with Individual Effects(NVARIE)model with individual effects can show the influence of different node response variables,it cannot show the social network in the spatial geographical structure.The Spatio temporal Dependent Social Network Model(STDSN)proposed in this paper takes into account the impact of spatial distance on node response variables,and sets a spatial correlation structure on the error items of the model.3.In this paper,the two-step estimation method is used to estimate the parameters more efficiently.This paper refers to the two-step estimation method of the spatial error model,and estimates the spatial effects and other parameters of the spatio-temporal dependent social network model(STDSN)using the maximum likelihood estimation and the generalized least squares estimation methods.Then,numerical simulation is carried out to compare and analyze the estimation performance of NVAR,NVARIE and STDSN.It is found that STDSN takes full account of the space-time dependence of the network and has better performance.Finally,by comparing the spatiotemporal dependent social network model(STDSN)with other models,the following research conclusions can be drawn:1.The electric energy trade network has significant time dependence,and the national electric energy trade is affected by the electric energy trade volume and trade network relationship over the years.2.The electric energy trade network has significant spatial dependence.The electric energy trade between countries in the world is affected by the electric energy trade between neighboring countries.In addition,the territorial area of a country also positively affects the electric energy trade volume.3.National development has a significant impact on the volume of electricity trade.Developed countries tend to have a larger volume of electricity trade,while countries with large import and export volumes of electricity usually have complex trade relations.