基于不同样本分区和模型参数的四川省粮食产量空间化研究

发布时间：2018-07-06 16:58

本文选题：粮食 + 模型　；参考：《中国农业资源与区划》2017年09期

【摘要】：[目的]农业经济统计数据受多种自然条件因素和社会经济条件因素影响,数据空间化难度很大。通过省级尺度粮食产量数据空间化的研究,探索提高农业经济统计数据空间化精度的方法及经验。[方法]以四川省为例,按照不同分区方案(全省不分区、分为5个综合农业分区),选择不同样本数据(县级粮食总产、县级平均粮食产量)为因变量,对应2种耕地类型(水田、旱地)面积数据为自变量,考虑2种模型参数(常数为0和不为0)拟合多元回归模型,对粮食数据进行空间化研究,并选择误差评价因子对空间化结果进行评价。[结果](1)不管采用县级粮食总产还是县平均粮食产量拟合,常数项设置为非0的,均是分区比不分区的结果精度要高;(2)不管采用县级粮食总产还是县平均粮食产量拟合,常数项设置为0的,均是分区比不分区的结果精度要高;(3)对比不分区和分区这2种方法,以县级粮食总产拟合的结果要比县级平均粮食产量拟合的结果精度要高。[结论]在省级粮食产量空间化时,以分县平均粮食产量为基础,划分土地利用类型、划分农业分区并且常数为0时拟合精度最高。在今后的研究中,有必要结合更多的影响因子进行粮食产量的空间化,以提高其数据重构精度。
[Abstract]:[objective] the agricultural economic statistical data are influenced by many natural and social economic factors, so it is very difficult to spatialize the data. Based on the research of spatialization of grain yield data in provincial scale, the methods and experiences of improving the spatial accuracy of agricultural economic statistics data are explored. [methods] taking Sichuan Province as an example, according to different zoning schemes (the province is not divided into 5 integrated agricultural zones), different sample data (total grain production at county level and average grain yield at county level) were selected as dependent variables. The area data of two types of cultivated land (paddy field and dry land) were considered as independent variables, and two model parameters (constant 0 and no 0) were considered to fit the multiple regression model, and the grain data were spatially studied. The error evaluation factor is selected to evaluate the spatial results. [results] (1) whether the total grain production at the county level or the average grain yield at the county level were fitted, the constant term was set to be non-zero, and the precision of the results was higher in the district than in the non-zonal area; (2) whether the total grain production at the county level or the average grain yield at the county level was fitted, When the constant term is set to 0, the precision of the result of the partition is higher than that of the one without the partition. (3) comparing the two methods, the fitting result of the total grain yield at the county level is higher than that of the average grain yield at the county level. [conclusion] when the grain yield is spatialized at provincial level, the land use type is divided on the basis of the average grain yield by county, and the fitting precision is the highest when the partition of agriculture is divided and the constant is 0. In order to improve the precision of data reconstruction, it is necessary to use more factors to spatialize grain yield in the future research.
【作者单位】：四川省农业科学院农业信息与农村经济研究所;四川省农业科学院区域农业发展研究中心;四川省农业科学院大数据中心;
【基金】：四川省财政创新能力提升工程“大数据背景下主要农产品数量安全预警分析技术研究”(2016GYSH-004) 四川省科技支撑计划软科学项目“基于精细区划的农产品气候品质认证评价指标体系研究”(2017ZR0045)
【分类号】：F326.11

【相似文献】