基于随机森林的耕地利用效率测度模型构建及其应用

doi:10.31497/zrzyxb.20190617

摘要/Abstract

摘要：

构建合适的量化分析模型是科学把握耕地利用状况及利用效率的基础性工作,可为制定合理有效的耕地资源管控政策,实现耕地利用与生态环境的协调发展提供决策依据。为了更准确地反映耕地利用系统的复杂性、动态性及差异性等特征,鉴于随机森林的基本思想,运用随机抽样Bootstrap法在合理构建分类树的基础上,构造了耕地利用效率测度的RF模型,进而以中国粮食主产区172个城市为例训练该模型并将其运用至2003-2015年的耕地利用效率测度中,同时将BP神经网络和熵权法作为对比验证其一致性、代表性和优越性。结果表明：（1）耕地利用效率测度的RF模型不受量纲限制,运行所需参数少,运算过程简化,能够较为精确地模拟各评价指标间的复杂联系,科学量化各评价指标对耕地利用效率的贡献。（2）对同一空间单元的效率值而言,RF>BPNN>EW,RF与BPNN所得效率值的总体分布格局相似,且均与EW的测度结果存在较大差异。（3）从评价结果与现实的匹配度和精度表征参数来看,RF的测度结果与自然和社会经济发展等客观事实更相符,具有较高的适用性与可靠性。同时,与其余两种常用模型相比,RF能够降低计算复杂度,提高训练效率,其测度结果的相关系数R为0.8685,M_RPD为2.3533,且具有最小M_MSE0.0174和M_MAE0.0211,更适用于复杂非线性特征的耕地利用效率研究。

关键词: 耕地利用效率, 随机森林, 粮食主产区

Abstract:

Setting up a suitable quantitative analysis model is a basic work for scientific grasp of cultivated land utilization efficiency and its distribution pattern, and can provide reasonable decision-making basis for sustainable utilization of cultivated land then realizing the coordinated development of cultivated resources and environment. In order to effectively describe the complexity, dynamics and heterogeneity characteristics of cultivated land use system, a random forest (RF) model for measuring cultivated land utilization efficiency is constructed by applying random sampling Bootstrap to build a classification tree reasonably. Then by taking 172 cities in the major grain producing areas of China as an example, the RF model was trained to measure the cultivated land utilization efficiency in 2003-2015 compared with Back Propagation Neural Network and Entropy weight to verify the consistency, representative and superiority of RF. The results show that: (1) RF model has fewer parameters and simpler implementation. It can simulate the complex relations among the evaluation indexes, which makes it convenient to analyze the value of each index. (2) For efficiency measurement results of the same space unit, RF > BPNN > EW, the overall distribution pattern of the cultivated land utilization efficiency in RF and BPNN is similar while a great difference exists in EW. (3) Judged from the matching degree of evaluation results to reality and the accuracy parameters, the measurement results are reasonable and in accordance with the facts in RF, which reflected its high applicability and reliability. At the same time, compared with the other two commonly used models, RF can reduce the dimensions of input vectors and the computing complexity, then raise the training efficiency. The correlation coefficient R of RF is 0.8685, M_RPD is 2.3533, with the minimum M_MSE and M_MAE being 0.0174 and 0.0211, respectively, which is more suitable for the study of the cultivated land utilization efficiency with complex nonlinear characteristics, and this method has explored a new way for evaluating cultivated land utilization efficiency.

Key words: cultivated land utilization efficiency, random forest, main grain producing areas

陈丹玲, 卢新海, 匡兵. 基于随机森林的耕地利用效率测度模型构建及其应用[J]. 自然资源学报, 2019, 34(6): 1345-1356.

Dan-ling CHEN, Xin-hai LU, Bing KUANG. Measurement of cultivated land utilization efficiency: Construction and application of random forest[J]. JOURNAL OF NATURAL RESOURCES, 2019, 34(6): 1345-1356.

图/表 10

图1

图2

图3

表1

表2

图4

表3

图5

图6

表4

参考文献 18

[1]	JIANG G ZHANG R, MA W, et al. Cultivated land productivity potential improvement in land consolidation schemes in Shenyang, China: Assessment and policy implications. Land Use Policy the International Journal Covering All Aspects of Land Use, 2017, 68: 80-88.
[2]	卢新海, 匡兵, 李菁. 碳排放约束下耕地利用效率的区域差异及其影响因素. 自然资源学报, 2018, 33(4): 657-668.
	[LU X H, KUANG B, LI J.Regional differences and its influencing factors of cultivated land use efficiency under carbon emission constraint. Journal of Natural Resources, 2018, 33(4): 657-668.]
[3]	KLEIJN D, KOHLER F, BALDI A, et al.On the relationship between farmland biodiversity and land-use intensity in Europe. Proceedings Biological Sciences, 2009, 276(1658): 903-909.
[4]	龙禹桥, 吴文斌, 余强毅, 等. 耕地集约化利用研究进展评述. 自然资源学报, 2018, 33(2): 337-350.
	[LONG Y Q, WU W B, YU Q Y, et al.Recent study progresses in intensive use of cropland. Journal of Natural Resources, 2018, 33(2): 337-350.]
[5]	杜国明, 刘彦随. 黑龙江省耕地集约利用评价及分区研究. 资源科学, 2013, 35(3): 554-560.
	[DU G M, LIU Y S.Evaluating and zoning intensive utilization of cultivated land in Heilongjiang province. Resources Science, 2013, 35(3): 554-560.]
[6]	曹银贵, 周伟, 王静, 等. 基于主成分分析与层次分析的三峡库区耕地集约利用对比. 农业工程学报, 2010, 26(4): 291-296.
	[CAO Y G, ZHOU W, WANG J, et al.Comparative on regional cultivated land intensive use based on principal component analysis and analytic hierarchy process in Three Gorges Reservoir Area. Transactions of the CSAE, 2010, 26(4): 291-296.]
[7]	WANG K, ZHANG P.The research on impact factors and characteristic of cultivated land resources use efficiency: Take Henan province, China as a case study. Ieri Procedia, 2013, 5(5): 2-9.
[8]	李强, 彭文英, 王建强, 等. 乡镇企业发达区耕地健康评价与驱动机理研究. 自然资源学报, 2015, 30(9): 1499-1510.
	[LI Q, PENG W Y, WANG J Q, et al.Health assessment and driving mechanism analysis of cultivated land in the township enterprises developed region. Journal of Natural Resources, 2015, 30(9): 1499-1510.]
[9]	石淑芹, 曹玉青, 吴文斌, 等. 耕地集约化评价指标体系与评价方法研究进展. 中国农业科学, 2017, 50(7): 1210-1222.
	[SHI S Q, CAO Y Q, WU W B, et al.Progresses in research of evaluation index system and its method on arable land Intensification: A review. Scientia Agricultura Sinica, 2017, 50(7): 1210-1222.]
[10]	MENG X L, SHI F G.An extended data envelopment analysis for the decision-making. Journal of Inequalities & Applications, 2017, 2017(1): 240.
[11]	赖红松, 吴次芳. 基于粗糙集和支持向量机的标准农田地力等级评价. 自然资源学报, 2011, 26(12): 2141-2154.
	[LAI H S, WU C F.Productivity evaluation of standard cultivated land based on rough set and support vector machine. Journal of Natural Resources, 2011, 26(12): 2141-2154.]
[12]	LEO B.Random forests. Machine Learning, 2001, 45(1): 5-32.
[13]	赖成光, 陈晓宏, 赵仕威, 等. 基于随机森林的洪灾风险评价模型及其应用. 水利学报, 2015, 46(1): 58-66.
	[LAI C G, CHEN X H, ZHAO S W, et al.A flood risk assessment model based on Random Forest and its application. Journal of Hydraulic Engineering, 2015, 46(1): 58-66.]
[14]	LINDNER C, BROMILEY P A, IONITA M C, et al.Robust and accurate shape model matching using Random Forest Regression-Voting. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2015, 37(9): 1862-1874.
[15]	刘影, 肖池伟, 李鹏, 等. 1978-2013年中国粮食主产区“粮—经”关系分析. 资源科学, 2015, 37(10): 1891-1901.
	[LIU Y, XIAO C W, LI P, et al.Relationship of grain output and economic development from 1978 to 2013 in the major grain producing area of China. Resources Science, 2015, 37(10): 1891-1901.]
[16]	张立新, 朱道林, 谢保鹏, 等. 中国粮食主产区耕地利用效率时空格局演变及影响因素: 基于180个地级市的实证研究. 资源科学, 2017, 39(4): 608-619.
	[ZHANG L X, ZHU D L, XIE B P, et al.Spatiotemporal pattern evolvement and driving factors of cultivated land utilization efficiency of the major grain producing area in China. Resources Science, 2017, 39(4): 608-619.]
[17]	倪超, 杨胜天, 罗娅, 等. 基于循环经济的黑龙江省耕地利用集约度时空差异. 地理研究, 2015, 34(2): 341-350.
	[NI C, YANG S T, LUO Y, et al.The spatial-temporal difference analysis of cultivated land use intensity in Heilongjiang province based on circular economy. Geographical Research, 2015, 34(2): 341-350.]
[18]	SONG X, ZHU O, LI Y, et al.Cultivated land use change in China, 1999-2007: Policy development perspectives. Journal of Geographical Sciences, 2012, 22(6): 1061-1078.

准则层	指标层	计算依据	说明
投入水平	地均农业机械投入/(kW·hm^-2)	农业机械总动力/耕地面积	保留
	地均科技投入/(10⁴元·hm^-2)	技术投入总额/耕地面积	删除
	地均劳动力投入/(人·hm^-2)	农业从业人数/耕地面积	保留
	地均化肥投入/(kg·hm^-2)	化肥施用量/耕地面积	保留
利用强度	垦殖指数/%	耕地面积/土地总面积^[9]	删除
	灌溉指数/%	耕地有效灌溉率	保留
	复种指数/%	农作物播种总面积/耕地面积^[5]	保留
	稳产指数/%	旱涝收保面积/耕地面积^[9]	保留
	农膜利用强度/(kg·hm^-2)	农膜使用量/耕地面积	删除
产出效益	粮食单产/(10⁴kg·hm^-2)	粮食总产量/耕地面积	保留
	农民人均农业产值/(10⁴元·人^-1)	农业产值/农业人口数	保留
	单位面积农业产值/(10⁴元·hm^-2)	农业产值/耕地面积	保留
	万元产值能耗/(kwh·元^-1)	农业用电量/农业总产值	删除
可持续性	人均耕地面积/(hm²·人^-1)	耕地面积/总人口	保留
	非农指数/%	非农业人口总数/总人口	删除
	粮食安全系数/(kg·人^-1)	人均粮食占有量/400 kg	保留
	农业自然灾害成灾率/%	—	保留
	森林覆盖率/%	—	删除

试验	BPNNW	RFW	试验	BPNNW	RFW
1	1.0000	1.0000	7	0.5000	0.9300
2	1.0000	1.0000	8	0.7500	0.7500
3	0.9300	1.0000	9	0.9300	1.0000
4	0.7500	1.0000	10	1.0000	0.7500
5	0.5000	1.0000	均值	0.8110	0.9430
6	0.7500	1.0000

检验	标准偏差	M_MSE	M_MAE	M_RPD	R	显著性
RF	0.1423	0.0174	0.1119	2.3533	0.8685	0.0000
BPNN	0.1629	0.0211	0.1644	1.6806	0.5722	0.1240