The Parameter Estimation And Application Of Spatio-temporal Expansion Models

Nowadays,with the frequent occurrence of massive data and the increasing complexity of the observed sample data generation mechanism,Economic,environmental,ecological,health and other areas of time and space data statistical analysis has been more and more attention.Space-time data has the following characteristics:In a time section,it belongs to spatial data,with spatial geography of the relevance;and in a particular geographical position,the observed value is a time series,with the time factor on the correlation.The statistical inference of spatiotemporal data is divided into two categories: parameter regression and nonparametric regression,nonparametric regression is characterized by exploratory,localized,combined with visualization results and the lack of strict statistical inference theory,the computational complexity.Therefore,this paper selects a parametric regression model with verifiability,overall and good explanatory.A comprehensive description of the expansion of the background and development of the method,inspired by this method,the temporal and spatial characteristics of the parameters are incorporated into the regression framework,and the temporal and spatial expansion model is proposed.Based on the Le Sage prior distribution and Bayesian estimation of multiple linear regression model,the parameter estimation method of this model is obtained.By estimating the posterior distribution of the model parameters,the parameters of the spatiotemporal expansion model are obtained by Gibbs sampling.From the absolute deviation of the parameters,the mean value of the standard deviation and the sum of squares of residuals are compared with the ordinary least squares estimates,further explain the validity of this model and the effectiveness of the estimation method.Through this empirical analysis,it is revealed that the spatial and temporal expansion model has a strong explanatory power for the parameters,and it proves the existence of temporal and spatial heterogeneity of sulfur dioxide emissions.Finally,we use the spatiotemporal expansion model to study an example of the atmospheric environment.We use space-time expansion model to explore temporal and spatial characteristics between urbanization and 2emission,and between energy consumption and 2emission in China according to China's inter-provincial spatio-temporal data between 2004–2014.Taking the regression model as the basic framework,the expansion of geographic position and time to parameters in the model is extended,so that we can obtain the time and space expansion model for analyzing the impact of urbanization and energy consumption on 2emission.The empirical analysis shows that,with the increase of urbanization level,China's sulfur dioxide emission is reduced as a whole under the combined effect of energy consumption in per unit GDP and gross,along with treatment measures.The influence of urbanization level on 2emission shows temporal and spatial heterogeneity,from the perspective of space,the impact of urbanization level on 2emission increases from south to north with increasing latitude;from the perspective of time,the trend of time is gradually decreasing in all regions.In addition,the increase in energy consumption as a whole will promote the emission of sulfur dioxide,which also shows the characteristic of temporal and spatial changes,so its degree of influence in different regions differs,that is,the effect decreases from south to north with increasing latitude,and the influential intensity of all regions' energy consumption to sulfur dioxide emissions gradually decreases with time.