Study Of The Village-level Economy Based On Spatial Autocorrelation

Geographical entities are not fully independent of each other. Tobler has pointed out in the firstgeography law: space entities are always interrelated, and the demonstration of correlation between theentities nearer to each other is especially obvious, so, in the study of different scales of regional issues,spatial correlations, especially the spatial autocorrelation and special heterogeneity, must be taken intoconsideration, namely, the attributes or the phenomenon of the adjacent positions under various scales shallbe studied from spatial thinking and perspective. The method of spatial autocorrelation analysis has alreadybeen widely applied to the fields of nature, economy and society, and become an important method toquantitatively study related problems involving space relation and an effective means to analyze spatialpattern.As small regional systems, village-level are the main carriers of agricultural economic activities,they are not independent of, but related to, each other, and they constantly and mutually transmit andexchange materials, energy, personnel, information, funds and etc with the external world, i.e. spatialinteractions occur between village-level, which cause the space neighbor effects and have far-reachinginfluence on the economic development of village-level. In the past, scholars mainly explored the economicdevelopment of village-level from the respects of industrial and agricultural development, systematicculture, economic development model, professionalization, modernization, and etc, while they neglectedthe influences of the spatial relationship between village-level on their economic development.With the support of the theories of spatial interaction, agglomeration-diffusion, spatialautocorrelation and etc, this paper carries out an exploratory spatial data analysis (ESDA) for the292administrative villages in Gongyi city focusing on space correlation measurement, and in combination withthe main economic data of villages in the statistical yearbooks of and Gongyi of2003,2004,2005and2006and the electronic maps of villages in Gongyi city, the paper establishes a database of spatialautocorrelation analysis. On this basis, taking structuring spatial weight matrix by the KNN standard (k)and the spatial autocorrelation analysis of the software GeoDA as the main instruments, the paper analyzesand estimates the space distribution features of the village-level economy of Gongyi city in the years 2003-2006from the aspects of global spatial autocorrelation (Moran I index), local spatial autocorrelation(Moran I scatter diagram and LISA chart) and spatio-temporal correlation (Moran I scatter diagram withdouble variables) of the village-level economic of Gongyi city. At the same time, the paper discoversspatial agglomeration and spatial abnormity, reveals the spatial dependence and the heterogeneity, namely,the spatial distribution status of village-level economy and the distribution characteristics demonstrated,and draws the following conclusions:(1) The analyzing results of global spatial autocorrelation shows that, the village-level economyof Gongyi city in2003-2006demonstrates a significant positive correlation, and presents a strong spatialagglomeration characteristics, that is, the development of village-level economy of Gongyi city is notcompletely in a random state, on the contrary, it indicates a spatial agglomeration of the similar values (thedeveloping levels of village-level economy are close). Besides, the spatial agglomeration in the studyperiod reveals a general trend of increasing first while declining later, and though the Moran value of2004is the largest, demonstrating the most significant spatial agglomeration, the overall spatial agglomerationlevel is higher.(2) The analysis of the estimating results of Moran scatter diagram and LISA chart reveals that,the village-level economy of Gongyi city in2003-2006demonstrates spatial dependency (spatialagglomeration) and heterogeneity (spatial abnormality). At the time it shows obviously “high-high”,“low-low” spatial agglomeration, it also shows “high-low” and “low-high” spatial abnormality. The overallspace distribution pattern is consistent, and it is relatively stable in the study period. On the basis of thediscovery of spatial heterogeneity, some information and phenomenon covered in global analysis is dug out,so that some special conditions of the local village-level (spatial abnormal points) are presented, so as toassist further analysis and the construction of related policies. At the same time, through the contrastiveanalysis of the different significant levels of LISA agglomeration graphs in LISA chart, it can be seen thatthe special conditions have great influences on the stability of the estimating results, and they cansignificantly change general pattern of the distribution of the spatial agglomeration.(3) Through analysis of the Moran scatter diagram with dual variables, the results show that thevillage-level economy of Gongyi city in2003-2006has a significant positive spatio-temporal correlation, displays a strong spatio-temporal agglomeration features, and that the spatio-temporal correlations ofvillage-level economy in2004and2005are the most significant.(4) The related analysis shows that, the ESDA analyzing method used for the related studies ofthe village-level economy both overcomes the shortcomings of the traditional method which emphasizes“sample independence”, and makes up for the characteristics which neglect sample’s spatial location andspatial correlation, that is, the traditional method only makes intuitive representation for the correspondingproperty values by way of special maps, so it lacks of concerns about the correlation characteristics of thevillage-level economy. However, in this paper, the research system is integrated into spatial dimensions,thus, quantitative analysis results are obtained so as to better reveal the characteristics of spatialdependence and heterogeneity of the village-level economy; simultaneously, the research scales of theanalysis method of spatial autocorrelation are deepened and its application fields are broadened.