Statistical Analysis Of Panel Data Of Poverty Accident In Guizhou Province

The purpose of the Statistics is descriptively and exploratory analysis the data variation to solve practical problems by the way of collecting and processing data for a specific research objectives There are three following mainly economic data structure forms:time series data、cross-sectional data and panel data. And the mainly statistical analysis methods is regression analysis,it means to effectively doing empirical analysis we should great different regression models depending on different data characteristics. In addition,Domestic and foreign scholars focus on the poverty recently,and their researches mainly focus on Measuring the extent of poverty、Calculating the depth of poverty、Exploring the distribution of poverty and Analysis the evolution of poverty. Since Office of the State Council Poverty Alleviation announced the newly Outline of Poverty from2011to2020, Local governments actively launch poverty works by the means of Contiguous development to Whole village. Guizhou province own the highest poverty rate,and also is a province featured as the largest area of poverty.、the deepest levels of poverty and the most important tasks of poor development,so government play highly attention to Guizhou Province poverty. In summery,this research aim at study Guizhou poverty by quantitative and qualitative analysis,then looking for the poverty factors,at last provide poverty reduce advices.At frist,do simply exploratory data analysis,then establish times series model for variable selection,at last build panel data model to understand on Individual and time diversification. This study mainly do following works:Firstly, the exploratory data analysis (EDA). Exploratory analysis on Head Count Ratio from2003to2011and88counties (city and district) over the years in Guizhou Province and chart analysis on poverty distribution.Second, the variable determination. Through comprehensive analysis on Guizhou Province situation, combining with economic, social and natural factors and choosing the suitable ones. Simultaneously, set up time-series model according to the Head Count Ratio from2003to2011(90samples) and implement optimizing selection about independent variable. Finally, it is to determine the variable.Third, the empirical analysis. To establish various panel data model on the basis of Head Count Ratio during2003to2011(2200samples) from88counties(city and district) and to define the final panel data model through F statistic test, AIC criterion and Pareto Optimality, then make a conclusion.Above all, this paper analyses the reasons for poverty in Guizhou on qualitative and quantitative those two aspects, extending poverty study from single measurement to quantitative analysis, spreading Head Count Ratio research from simple time-series model to panel data model, refining the poverty study from nationwide or province to counties (city and district).