基于密度峰值聚类的我国粮食产量的影响因素分析

摘要

为粮食增产、粮食安全及可持续发展提供理论依据,采用密度峰值聚类法对我国31个省(市、区)2000-2014年的粮食产量及其影响因素的面板数据进行聚类,同时引入变截距双对数模型对每个类簇进行计量分析.结果表明:农业发展水平不一致地区其影响粮食产量因素存在较大差别.除粮食播种面积、化肥施用量和受灾面积3个主要的共同影响因素外,农业发达地区、农业欠发达地区和农业不发达地区的次主要影响因素分别是农业机械总动力、畜牧业发展及农业劳动力.%The panel data of grain yield and influencing factors from 31 provinces during 2000～2014 were analyzed by the density peak clustering analysis and each class cluster was econometrically analyzed by the introduced variable intercept double logarithmic model to provide the theoretical basis for grain increase,grain security and grain sustainable development.Results:There are obvious differences in influencing grain yield factors in areas with a discord agricultural development level.Three main factors influencing grain yield are grain acreage,chemical fertilizer application rate and covered disaster area.The secondary main factor influencing grain yield in agricultural developed regions,agricultural less developed regions and agricultural underdevelopment regions is total power of agricultural machinery,animal husbandry development and agricultural labor force respectively.