基于偏最小二乘回归模型的湖北省粮食产量影响因素分析  

Analysis of Factors Influencing Grain Production of Hubei Province Based on Partial Least Squares Regression

作　　者：吴红霞[1] 李荆荆[2] 聂艳[1] 童秋英[3] 吴西子 

机构地区：[1]华中师范大学城市与环境科学学院,武汉430079 [2]湖北省国土资源研究院,武汉430071 [3]武汉市国土资源与规划信息中心,武汉430014

出　　处：《湖北农业科学》2017年第13期2553-2559,共7页Hubei Agricultural Sciences

基　　金：湖北省国土资源厅行业科研专项(ETZ2015A09);华中师范大学中央高校基本科研业务费专项资金(CCNU15A02049)

摘　　要：以湖北省为研究区域,选取9个影响粮食产量因素指标,通过对自变量之间的多重共线性进行分析诊断,构建了基于C-D生产函数的2000~2014年湖北省粮食产量影响因素的偏最小二乘回归(Partial least squares regression,PLSR)模型。PLSR模型中,自变量对因变量均具有较好的解释能力,回归模型的Rcv2=0.946,表明回归模型的精度较高,拟合效果较好,可靠性强。结果表明,粮食作物播种面积、农业机械化总动力、农田有效灌溉面积、农用化肥施用量以及农村用电量共5个指标是影响湖北省粮食产量的关键因素。Taking Hubei Province as the research area, 9 indicators were selected which affected grain production from 2000 to 2014 in Hubei province, the influence factors of grain production partial least-square regression model were constructed which based on C-D production function after analyzing and diagnosing the multiple mutual linear between the independent variables. The PLSR model of grain production indicated that the independent variables had good explanatory power for the dependent variables. Moreover, the goodness of fit Rcv2=0.946 of regression model showed there were high precision, strong reliability and better fitting effect. The results indicated that 5 indicators including the grain sown area, the total power of agricultural machinery, the effective irrigated area of farmland, the quantity of chemical fertilizer and rural power consumption were the main factors which affected grain yield in Hubei Province.

关 键 词：偏最小二乘回归分析 粮食产量 影响因素 湖北省 

分 类 号：F30[经济管理—产业经济]