Research On Spatial Autocorrelation Analysis Method And Its Application

The Application of spatial statistics is the new hot spot in today’s statistical frontier. Application of the new focus of the theses in this field developed in recent years, namely spatial autocorrelation analysis. We first put the theory of spatial autocorrelation and improved the spatial and temporal autocorrelation indices, temporal and spatial autoregression model is applied to the field of marine ecosystems. We first introduced the space since the main developments related concepts, theories, models; and use of monitoring data in Wuhan Environmental Protection Monitoring the impact point on the spatial and temporal distribution of PM2.5 conducted data analysis and research, and the use of space autoregression analysis and temporal autoregressive model to study the four variables (SO2, NO2, CO, O3) temporal influence in January 2013 to August of PM2.5 distribution.In four other explanation for PM2.5 and vector (SO2, NO2, CO, O3) in the modeling process, we exponential model for temporal autocorrelation and temporal autoregressive model are explored. In the field of spatial statistics, application time and space in the current model is still a challenge. Spacetime Lopez-Herndndez and Chasco-Yrigoyen (2007) proposed since the foundation of the underlying index, we propose a high-order temporal and spatial autocorrelation index model and put it for the visualization in three-dimensional space. In order to consider the time-step directions autocorrelation, spatial economy on the basis of spatial data modeling process of Anselin given, we propose a space-time autoregressive model of PM2.5.The results show that: PM2.5 there are significant temporal and spatial autocorrelation, spatial and temporal distribution of PM2.5 undergone significant changes and volatility, and the spatial distribution of PM2.5 such changes and fluctuations something abnormal changes occur. Because of the space distribution of PM2.5 autocorrelation exists, making the ordinary linear regression model is no longer valid. In the third chapter, we use the residual space autoregression model (including spatial and temporal model model), the model can effectively solve the spatial distribution of PM2.5-related issues since.