基于GIS的抽样框编制与抽样技术方法研究
|
摘要
随着我国社会经济的快速发展,以周期性普查为基础、经常性抽样调查为主体、必要的统计报表、重点调查和综合分析为补充的统计调查业务也随之快速地发展,统计数据和统计信息量以几何级数迅猛增长。与此同时,国家统计部门、政府决策部门和社会公众对统计工作也提出了更高的要求——需要统计信息和空间信息资源的全面整合,需要全面、深入的统计分析和更及时、更完善的统计服务。这样,一些传统的抽样框组织管理和抽样技术方法也将随之发生变化和改进,以适应更加全面地需求。
本文利用地理信息系统技术,采用空间数据分析方法,将传统的抽样思路与空间信息相结合,提出了基于空间的“动态抽样框”编制技术,进行了空间分层抽样技术的研究利应用,并最终以北京市农村区域为例进行全市人均农业总产值的抽样调查。论文的主要研究内容和结论为:
(1)对抽样单元的空间自相关分析。空间自相关是空间地理数据的重要性质,依据对抽样单元的属性的自相关分析,来确定其空间上近邻的面积单元中该地理属性的相似性特征,最终判定抽样单元在空间中的集聚模式。文中对北京市人均农业总产值进行空间自相关分析,结果表明:北京市乡镇农业经济存在正的相关性特性,表现出很强的空间集聚特征。
(2)“动态抽样框”的编制。传统的抽样框是将抽样总体按照一定规则转化为抽样框,并对抽样单元进行编号,形成一个名录,因此该抽样框是固定形式,一经生成就不能再变化。这里,我们研究的是一种“动态”的抽样框模型,其抽样框的组成随着空间分析的变化而生成,并可对其随时进行调整。文中通过对人均农业总产值的空间分析,根据分析结果分为“高高分布”抽样框、“低低分布”抽样框、“高低分布”抽样框、“低高分布”抽样框和“分布不明显”抽样框,合理且准确。
(3)样本容量的估算与分配。抽样的核心是从抽样框中抽取一定量的样本,进而对样本进行分析用来推断总体。在这里我们对样本容量的估算和分配是基于空间意义上的,从样本单元空间分布的角度出发,依据样本单元的相似性进行样本数量的分配和抽取。根据样本容量的计算公式,在“市抽乡”的抽样过程中,从186个乡镇中抽取的样本数量为65个。根据样本的分配比例,每个抽样框的样本分配数量为:“高高分布”抽样框(14个)、“低低分布”抽样框(16个)、“高低分布”抽样框(4个)、“低高分布”抽样框(2个)和“分布不明显”抽样框(29个)。
(4)空间分层抽样方法的研究与应用。将传统的抽样方法与GIS(地理信息系统)相结合,并整合上述的一切关键技术,并最终对空间抽样的误差和准确度进行分析,从而形成空间抽样的一整套流程。其中,本文针对北京市人均农业总产值的抽样调查结果为:以±0.3861万元为抽样误差的取值区间(3.19,3.97)万元。
With the rapid social economic development of our country, the statistical survey services which based on the cyclical census and regular sampling surveys also develop rapidly; statistical data and statistical information grow by geometric series. At the same time, the government of statistics and decision-making department and the public also set a higher request to the statistical work. It needs the integration of statistical information and the spatial data, the deeply and comprehensive analysis of statistical analysis and more timely and improvement statistic services. So, some tradition's sampling methods and techniques need change to adapt more comprehensive requests.
Based on the Geographical Information System techniques and spatial data analysis methods, we combined the traditional sampling methods with spatial information in this article, proposed the "Dynamic Sampling Frames" concept, and the space stratified sampling techniques. At the end of this article, we take the Beijing rural areas as an example, to research and do some applications in the agricultural total output value per capita sampling surveys. The main contents are the followings:
(1) Spatial autocorrelation analysis of sampling unit. Spatial autocorrelation is an important nature of spatial geographical data. According to the autocorrelation analysis with sampling unit, the similar character of the neighbor unit and the combined model of sampling unit in the space can be can be identified. This paper takes a spatial autocorrelation analysis for agricultural GDP per capita in Beijing. The results showed that there is a positive correlation in Beijing township agricultural economy and a strong performance characteristic of the spatial concentration.
(2) The establishment of dynamic sampling frames. Traditional sampling frames are converted directly from the overall sampling, and number the sampling unit to a list. This kind of sampling frame is fixed and can not be changed again. This paper studies a dynamical sampling frame which is consisted of the spatial analysis change and could be adjusted at any moment.
(3) The estimation and distribution of sample capacity. The core of sampling is extracting a certain amount of samples from sample frames, and then inferring the overall through the analysis to samples. In this paper we estimate the sample capacity and distribution based on the sense of space. From the perspective of spatial distribution, we carried on the allocate and extraction of samples according to the similarity of samples' certain attribute.
(4) The research of space stratified sampling techniques. This technique combines the traditional sampling method with GIS, integration the above mentioned key techniques, and then analyze the spatial sampling error and accuracy, thus forming a complete spatial sampling process.
本文利用地理信息系统技术,采用空间数据分析方法,将传统的抽样思路与空间信息相结合,提出了基于空间的“动态抽样框”编制技术,进行了空间分层抽样技术的研究利应用,并最终以北京市农村区域为例进行全市人均农业总产值的抽样调查。论文的主要研究内容和结论为:
(1)对抽样单元的空间自相关分析。空间自相关是空间地理数据的重要性质,依据对抽样单元的属性的自相关分析,来确定其空间上近邻的面积单元中该地理属性的相似性特征,最终判定抽样单元在空间中的集聚模式。文中对北京市人均农业总产值进行空间自相关分析,结果表明:北京市乡镇农业经济存在正的相关性特性,表现出很强的空间集聚特征。
(2)“动态抽样框”的编制。传统的抽样框是将抽样总体按照一定规则转化为抽样框,并对抽样单元进行编号,形成一个名录,因此该抽样框是固定形式,一经生成就不能再变化。这里,我们研究的是一种“动态”的抽样框模型,其抽样框的组成随着空间分析的变化而生成,并可对其随时进行调整。文中通过对人均农业总产值的空间分析,根据分析结果分为“高高分布”抽样框、“低低分布”抽样框、“高低分布”抽样框、“低高分布”抽样框和“分布不明显”抽样框,合理且准确。
(3)样本容量的估算与分配。抽样的核心是从抽样框中抽取一定量的样本,进而对样本进行分析用来推断总体。在这里我们对样本容量的估算和分配是基于空间意义上的,从样本单元空间分布的角度出发,依据样本单元的相似性进行样本数量的分配和抽取。根据样本容量的计算公式,在“市抽乡”的抽样过程中,从186个乡镇中抽取的样本数量为65个。根据样本的分配比例,每个抽样框的样本分配数量为:“高高分布”抽样框(14个)、“低低分布”抽样框(16个)、“高低分布”抽样框(4个)、“低高分布”抽样框(2个)和“分布不明显”抽样框(29个)。
(4)空间分层抽样方法的研究与应用。将传统的抽样方法与GIS(地理信息系统)相结合,并整合上述的一切关键技术,并最终对空间抽样的误差和准确度进行分析,从而形成空间抽样的一整套流程。其中,本文针对北京市人均农业总产值的抽样调查结果为:以±0.3861万元为抽样误差的取值区间(3.19,3.97)万元。
With the rapid social economic development of our country, the statistical survey services which based on the cyclical census and regular sampling surveys also develop rapidly; statistical data and statistical information grow by geometric series. At the same time, the government of statistics and decision-making department and the public also set a higher request to the statistical work. It needs the integration of statistical information and the spatial data, the deeply and comprehensive analysis of statistical analysis and more timely and improvement statistic services. So, some tradition's sampling methods and techniques need change to adapt more comprehensive requests.
Based on the Geographical Information System techniques and spatial data analysis methods, we combined the traditional sampling methods with spatial information in this article, proposed the "Dynamic Sampling Frames" concept, and the space stratified sampling techniques. At the end of this article, we take the Beijing rural areas as an example, to research and do some applications in the agricultural total output value per capita sampling surveys. The main contents are the followings:
(1) Spatial autocorrelation analysis of sampling unit. Spatial autocorrelation is an important nature of spatial geographical data. According to the autocorrelation analysis with sampling unit, the similar character of the neighbor unit and the combined model of sampling unit in the space can be can be identified. This paper takes a spatial autocorrelation analysis for agricultural GDP per capita in Beijing. The results showed that there is a positive correlation in Beijing township agricultural economy and a strong performance characteristic of the spatial concentration.
(2) The establishment of dynamic sampling frames. Traditional sampling frames are converted directly from the overall sampling, and number the sampling unit to a list. This kind of sampling frame is fixed and can not be changed again. This paper studies a dynamical sampling frame which is consisted of the spatial analysis change and could be adjusted at any moment.
(3) The estimation and distribution of sample capacity. The core of sampling is extracting a certain amount of samples from sample frames, and then inferring the overall through the analysis to samples. In this paper we estimate the sample capacity and distribution based on the sense of space. From the perspective of spatial distribution, we carried on the allocate and extraction of samples according to the similarity of samples' certain attribute.
(4) The research of space stratified sampling techniques. This technique combines the traditional sampling method with GIS, integration the above mentioned key techniques, and then analyze the spatial sampling error and accuracy, thus forming a complete spatial sampling process.
引文
[1].Cochran W G.Sampling Techniques.3d ed.New York:John Wiley & Sons,1977.1-355.
[2].Kish L.Survey Sampling.New York:John Wiley & Sons,1985.1-327,644-691.
[3].李金昌.应用抽样技术[M].科学出版社,2006.
[4].曾玉平,张勇.中国农产量抽样调查的发展历程.中国统计,2004,(3):14-16.
[5].庄大方,张稳,罗建国.土地资源遥感调查中的空间信息多重采样框架设计与GIS实现[J].统计研究,1999(1):37-41.
[6].庄大方,刘明亮,罗建国.遥感和空间分析支持下的耕地、城镇动态变化空间采样方法研究[J].统计研究,1999(10):45-5.
[7].徐新良等.基于3S技术的土地利用/土地覆盖变化野外采样框设计—以东北地区黑龙江省为例[J].遥感技术与应用,2002,17(3):135-139.
[8].刘海启.大尺度耕地变化监测的遥感抽样方法研究[J].农业工程学报,2001,17(2):168-171.
[9].李连发,王劲峰,刘纪远.国土遥感调查的空间抽样优化决策[J].中国科学(D辑)地球科学,2004,34(10):975-982.
[10].黄敬峰等.基于GIS的水稻遥感估产模型研究[J].遥感技术与应用,2002,17(3):125-128.
[11].焦险峰等.全国棉花种植面积遥感监测抽样方法设计[J].农业工程学报,2002.18(4):159-162.
[12].彭正强,朱朝华.腰果云翅斑螟幼虫空间分布型及抽样方法的研究[J].华南农业大学学报(自然科学版),2002,23(3):43-48.
[13].罗于洋等.柠条种子害虫幼虫空间分布型及抽样技术研究[J].西北林学院学报,2006,21(2):97-99.
[14].Ignacio R I,Jose M M.The design of rainfall networks in time and space.Water Resources Reaearch,1974,10:713.
[15].Haining R.Estimating spatial means with an application to remote sensing data.Communication Statistics Theory Meth,1988,17(2):537.
[16].Atkinson P M,Emery D R.Exploring the relation between spatial structure and wavelength:Implications for sampling reflectance in the field.International Journal of Remote Sensing,1999,20:2663.
[17].Atkinson P M.Geographical information science:Geostatistics and Uncertainty.Process in Physical Geography,1999,23:134.
[18].Journel A,Huijbregts C H.Mining geostatistics.London:Academic Press Inc..1978.
[19].Wang Jinfeng,Liu J,Zhuang D et al.Spatial sampling design for monitoring the area of cultivated land.International Journal of Remote Sensing,2002,13(2):263-284.
[20].A.Unwin,D.Unwin.Exploratory Spatial Data Analysis with Local Statistics[J].The Statistician,1998,47(3):415-421.
[21].陈斐,高志刚等.新疆各县市空间经济关联分析初步研究[J].干旱区地理,2003,26(3):274-279.
[22].胡霞,魏作磊.中国城市服务业发展差异的空间经济计量分析[J].统计研究,2006,(9):54-58.
[23].陈斐,陈秀山.局部空间统计在区域经济分析中的应用[J].华中师范大学学报(人文社会科学 版),2006,45(4):51-55.
[24].徐建华,岳文泽,谈文琦.城市景观格局尺度效应的空间统计规律[J].地理学报,2004,59(6):1058-1067.
[25].潘竟虎,张佳龙,张勇.甘肃省区域经济空间差异的ESDA-GIS分析[J].西北师范大学学报(自然科学版),2006,(06):67-70.
[26].吴玉鸣,李建霞.中国区域工业全要素生产率的空间计量经济分析[J].地理科学,2006,26(4):385-391.
[27].韦素琼,张金前,陈健飞.基于空间自相关的闽台城镇建设用地分布研究[J].地理科学进展,2007,26(3):12-17.
[28].吴玉鸣,徐建华.中国区域经济增长集聚的空间统计分析[J].地理科学,2004,24(6):654-659.
[29].冯士雍,施锡铨.抽样调查-理论、方法与实践[M],上海科学技术出版社,1996.
[30].Kish L著,倪加勋译.抽样调查[M].中国统计出版社,1997.
[31].赵宪文.林业遥感定量估测.中国林业出版社,1997:1-300.
[32].吴炳方.全国农情监测与估产运行化遥感方法.地理学报,2000,55(1):25.
[33].王劲峰等.空间分析[M].科学出版社,2006.
[34].李连发,王劲峰.地理数据空间抽样模型.自然科学进展,2005,12(05):545-548.
[35].王劲峰等.地理空间信息分析的理论体系探讨.地理学报,2000,55(1):92.
[36].崔国平.如何编好抽样框.统计与信息论坛,2003,18(1):80-80.
[37].李小建,乔家君.20世纪90年代中国县际经济差异的空间分析.地理学报,2001,56(2):136-145.
[38].郭庆旺,贾俊雪.中国区域经济趋同与差异的因素贡献分析[J].财贸经济,2006(2):11-17.
[39].胡良民,苗长虹,乔家君.河南省区域经济发展差异及其时空格局研究[J].地理科学进展,2002,21(3):268-274.
[40].欧向军,陈修颖.改革开放以来江苏省区域经济差异成因分析[J].经济地理,2004,24(3):338-343.
[41].陈斐.具有空间统计与GIS的区域经济分析研究[博士学位论文].武汉:武汉大学.2002.
[42].Goodchild M,Anselin L,Applebaum R,et al.Towards a spatial integrated social Science[J].International Regional Science Review,2000,23:139-159.
[43].陈斐,郭朝辉等.基于GIS的区域经济分析与决策初步研究[J].人文地理,2007,17(6):76-80.
[44].孟斌,王劲峰,张文忠.基于空间分析方法的中国区域经济差异研究[J].地理科学,2005,25(4):393-400.
[45].何江,张馨之.中国区域人均GDP增长速度的探索性空间数据分析[J].统计与决策,2006,72-74.
[46].鲁凤,徐建华.中国区域经济差异的空间统计分析[J].华东师范大学学报(自然科学版),2007,(2):45-51.
[47].樊新生,李小建.基于县域尺度的经济增长空间自相关性研究—以河南省为例[J].经济经纬,2005,(3):57-60.
[48].陈建建,甄峰.20世纪90年代以来安徽省区域经济差异时空特征分析[J].安徽农业科学,2007,35(9):2633-2635.
[49].Anselin L.Interactive techniques and exploratory spatial data analysis.In:Longley P A,Goodchild M F,Maguire D J,et al.Geographical Information Systems,Principles,Technical Issues,Management Issues and Applications.John Wiley & Sons,Inc,1999.253-266.
[50].Cliff A D,Ord J K.Spatial Autocorrelation[M].Pion,London,1973.
[51].Sokal R R,Oden N L.Spatial Autocorrelation in Biology,1.Methodology.Biological Journal of the Linnean Society,1978a,10,199-228.
[52].Goodchild M F.Spatial Autocorrelation(CATMOG 47).Norwich,UK:GeoBooks,1986.
[53].Cliff A D,Ord J K.Spatial Processes:Models and Applications[M].Pion,London,1981.
[54].马荣华,蒲英霞,马晓冬.GIS空间关联模式发现[M].科学出版社,北京,2007.
[55].农村抽样调查网点抽选方案(试行).国家统计局,1984.
[56].北京市统计局,国家统计局北京调查总队.北京农村统计资料(2005年暨“十五”时期).
[57].王远飞,何洪林.空间数据分析方法[M].科学出版社,北京,2007.
[2].Kish L.Survey Sampling.New York:John Wiley & Sons,1985.1-327,644-691.
[3].李金昌.应用抽样技术[M].科学出版社,2006.
[4].曾玉平,张勇.中国农产量抽样调查的发展历程.中国统计,2004,(3):14-16.
[5].庄大方,张稳,罗建国.土地资源遥感调查中的空间信息多重采样框架设计与GIS实现[J].统计研究,1999(1):37-41.
[6].庄大方,刘明亮,罗建国.遥感和空间分析支持下的耕地、城镇动态变化空间采样方法研究[J].统计研究,1999(10):45-5.
[7].徐新良等.基于3S技术的土地利用/土地覆盖变化野外采样框设计—以东北地区黑龙江省为例[J].遥感技术与应用,2002,17(3):135-139.
[8].刘海启.大尺度耕地变化监测的遥感抽样方法研究[J].农业工程学报,2001,17(2):168-171.
[9].李连发,王劲峰,刘纪远.国土遥感调查的空间抽样优化决策[J].中国科学(D辑)地球科学,2004,34(10):975-982.
[10].黄敬峰等.基于GIS的水稻遥感估产模型研究[J].遥感技术与应用,2002,17(3):125-128.
[11].焦险峰等.全国棉花种植面积遥感监测抽样方法设计[J].农业工程学报,2002.18(4):159-162.
[12].彭正强,朱朝华.腰果云翅斑螟幼虫空间分布型及抽样方法的研究[J].华南农业大学学报(自然科学版),2002,23(3):43-48.
[13].罗于洋等.柠条种子害虫幼虫空间分布型及抽样技术研究[J].西北林学院学报,2006,21(2):97-99.
[14].Ignacio R I,Jose M M.The design of rainfall networks in time and space.Water Resources Reaearch,1974,10:713.
[15].Haining R.Estimating spatial means with an application to remote sensing data.Communication Statistics Theory Meth,1988,17(2):537.
[16].Atkinson P M,Emery D R.Exploring the relation between spatial structure and wavelength:Implications for sampling reflectance in the field.International Journal of Remote Sensing,1999,20:2663.
[17].Atkinson P M.Geographical information science:Geostatistics and Uncertainty.Process in Physical Geography,1999,23:134.
[18].Journel A,Huijbregts C H.Mining geostatistics.London:Academic Press Inc..1978.
[19].Wang Jinfeng,Liu J,Zhuang D et al.Spatial sampling design for monitoring the area of cultivated land.International Journal of Remote Sensing,2002,13(2):263-284.
[20].A.Unwin,D.Unwin.Exploratory Spatial Data Analysis with Local Statistics[J].The Statistician,1998,47(3):415-421.
[21].陈斐,高志刚等.新疆各县市空间经济关联分析初步研究[J].干旱区地理,2003,26(3):274-279.
[22].胡霞,魏作磊.中国城市服务业发展差异的空间经济计量分析[J].统计研究,2006,(9):54-58.
[23].陈斐,陈秀山.局部空间统计在区域经济分析中的应用[J].华中师范大学学报(人文社会科学 版),2006,45(4):51-55.
[24].徐建华,岳文泽,谈文琦.城市景观格局尺度效应的空间统计规律[J].地理学报,2004,59(6):1058-1067.
[25].潘竟虎,张佳龙,张勇.甘肃省区域经济空间差异的ESDA-GIS分析[J].西北师范大学学报(自然科学版),2006,(06):67-70.
[26].吴玉鸣,李建霞.中国区域工业全要素生产率的空间计量经济分析[J].地理科学,2006,26(4):385-391.
[27].韦素琼,张金前,陈健飞.基于空间自相关的闽台城镇建设用地分布研究[J].地理科学进展,2007,26(3):12-17.
[28].吴玉鸣,徐建华.中国区域经济增长集聚的空间统计分析[J].地理科学,2004,24(6):654-659.
[29].冯士雍,施锡铨.抽样调查-理论、方法与实践[M],上海科学技术出版社,1996.
[30].Kish L著,倪加勋译.抽样调查[M].中国统计出版社,1997.
[31].赵宪文.林业遥感定量估测.中国林业出版社,1997:1-300.
[32].吴炳方.全国农情监测与估产运行化遥感方法.地理学报,2000,55(1):25.
[33].王劲峰等.空间分析[M].科学出版社,2006.
[34].李连发,王劲峰.地理数据空间抽样模型.自然科学进展,2005,12(05):545-548.
[35].王劲峰等.地理空间信息分析的理论体系探讨.地理学报,2000,55(1):92.
[36].崔国平.如何编好抽样框.统计与信息论坛,2003,18(1):80-80.
[37].李小建,乔家君.20世纪90年代中国县际经济差异的空间分析.地理学报,2001,56(2):136-145.
[38].郭庆旺,贾俊雪.中国区域经济趋同与差异的因素贡献分析[J].财贸经济,2006(2):11-17.
[39].胡良民,苗长虹,乔家君.河南省区域经济发展差异及其时空格局研究[J].地理科学进展,2002,21(3):268-274.
[40].欧向军,陈修颖.改革开放以来江苏省区域经济差异成因分析[J].经济地理,2004,24(3):338-343.
[41].陈斐.具有空间统计与GIS的区域经济分析研究[博士学位论文].武汉:武汉大学.2002.
[42].Goodchild M,Anselin L,Applebaum R,et al.Towards a spatial integrated social Science[J].International Regional Science Review,2000,23:139-159.
[43].陈斐,郭朝辉等.基于GIS的区域经济分析与决策初步研究[J].人文地理,2007,17(6):76-80.
[44].孟斌,王劲峰,张文忠.基于空间分析方法的中国区域经济差异研究[J].地理科学,2005,25(4):393-400.
[45].何江,张馨之.中国区域人均GDP增长速度的探索性空间数据分析[J].统计与决策,2006,72-74.
[46].鲁凤,徐建华.中国区域经济差异的空间统计分析[J].华东师范大学学报(自然科学版),2007,(2):45-51.
[47].樊新生,李小建.基于县域尺度的经济增长空间自相关性研究—以河南省为例[J].经济经纬,2005,(3):57-60.
[48].陈建建,甄峰.20世纪90年代以来安徽省区域经济差异时空特征分析[J].安徽农业科学,2007,35(9):2633-2635.
[49].Anselin L.Interactive techniques and exploratory spatial data analysis.In:Longley P A,Goodchild M F,Maguire D J,et al.Geographical Information Systems,Principles,Technical Issues,Management Issues and Applications.John Wiley & Sons,Inc,1999.253-266.
[50].Cliff A D,Ord J K.Spatial Autocorrelation[M].Pion,London,1973.
[51].Sokal R R,Oden N L.Spatial Autocorrelation in Biology,1.Methodology.Biological Journal of the Linnean Society,1978a,10,199-228.
[52].Goodchild M F.Spatial Autocorrelation(CATMOG 47).Norwich,UK:GeoBooks,1986.
[53].Cliff A D,Ord J K.Spatial Processes:Models and Applications[M].Pion,London,1981.
[54].马荣华,蒲英霞,马晓冬.GIS空间关联模式发现[M].科学出版社,北京,2007.
[55].农村抽样调查网点抽选方案(试行).国家统计局,1984.
[56].北京市统计局,国家统计局北京调查总队.北京农村统计资料(2005年暨“十五”时期).
[57].王远飞,何洪林.空间数据分析方法[M].科学出版社,北京,2007.