Research On Grey Incidence Models For Panel Data And Their Applications

Increasingly serious air pollution has been seriously affecting people's production,life,physical and mental health,and transportation,and it has already become a major environmental and social problem in the whole world.China's air pollution has typical regional and complex feature,which leads to the result that traditional one-dimensional data analysis methods have limitations in dealing with such problems.Considering the grey characteristic of short time series and limited data volume of air pollution monitoring data,the grey incidence model for the relationship between indicators of panel data and the grey incidence model for the relationship between objects of panel data are proposed respectively.Then,the proposed models are applied into the regional division and the recognition of main influencing factors of air pollution control in the Yangtze River Delta region.The specific research contents and conclusions are as follows:(1)The new grey incidence operators for panel data are proposed.Combining the dimensional characteristics of the panel data,the spatial characterization and matrix characterization of panel data are put forward.Based on which,the new grey averaging operator,standard operator,interval operator and logarithmic operator for panel data is constructed.Then,the basic characteristics of above several classes of operators are analyzed.Finally,the properties of several classes of operators,such as multiple transformation,translational transformation,idempotence and integrity,are discussed.(2)A novel grey incidence model for the relationship between indicators of panel data is constructed.Based on the difference between the time dimension and the object dimension of panel data,increment and deviation are introduced to respectively represent development level of indicator in the time dimension and distribution characteristics of indicator in the object dimension,and the direction difference between the two dimensions is used as the positive and negative correlation.Then,the difference in directions between the above two dimensions is introduced to be a judgement standard of positive and negative incidence,and the incidence coefficients of the time dimension and object dimension are built.Moreover,the grey indicator incidence model of panel data is constructed and the uniqueness,symmetry and comparability of this model are verified.(3)A novel grey incidence model for the relationship between objects of panel data is constructed.Based on the difference between the time dimension and the object dimension of panel data,growth rate and percentage difference are introduced to respectively represent growth speed of indicator in the time dimension and approaching level of object in the indicator dimension.Then,the incidence coefficients of the time dimension and object dimension are built.Moreover,the grey object incidence model of panel data is constructed.Finally,the critical values of degree of grey incidence is set and used to cluster the objects.(4)The regional division and the main influencing factors analysis of air pollution control in the Yangtze River Delta region.1)Firstly,the air pollution status in the Yangtze River Delta region is analyzed.Then,based on the spatial and temporal characteristics of the six main types of air pollutants,the grey object incidence model of panel data is employed to the regional division of pollution synergy in the Yangtze River Delta region,and this region is divided to 6 categories.2)From the macroeconomic perspective,the sources of atmospheric pollution in the Yangtze River Delta are summarized into five aspects: agriculture,industry,building,transportation and resident life.The proposed grey indicator incidence model for panel data is used to analyze air quality and pollution source for the relationship between indicators,and identify the main influencing factors of various classes of regions.Finally,based on the model results and analysis conclusions,the countermeasures and suggestions are put forward from the aspects: source control and industrial optimization.Research on Grey Incidence Models for Panel Data and their Applications.