基于随机森林的元胞自动机城市扩展模拟——以佛山市为例

doi:10.18306/dlkxjz.2015.08.001

摘要/Abstract

摘要：

本文提出一种基于随机森林的元胞自动机城市扩展(RF-CA)模型。通过在多个决策树的生成过程中分别对训练样本集和分裂节点的候选空间变量引入随机因素,提取城市扩展元胞自动机的转换规则。该模型便于并行构建,能在运算量没有显著增加的前提下提高预测的精度,对城市扩展中存在的随机因素有较强的容忍度。RF-CA模型可进行袋外误差估计,以快速获取模型参数;也可度量空间变量重要性,解释各空间变量在城市扩展中的作用。将该模型应用于佛山市1988-2012年的城市扩展模拟中,结果表明,与常用的逻辑回归模型相比,RF-CA模型进行模拟和预测分别能够提高1.7%和2.6%的精度,非常适用于复杂非线性特征的城市系统演变模型与扩展研究;通过对影响佛山市城市扩展的空间变量进行重要性度量,发现对佛山城市扩张模拟研究而言,距国道的距离与距城市中心的距离具有最重要的作用。

关键词: 随机森林, 元胞自动机, 城市扩展, 佛山

Abstract:

Cellular automata (CA) has been frequently used to investigate the logical nature of self-reproducible systems and simulate the evolution of complex geographical phenomena such as urban expansion. The core of cellular automata is to define transition rules. Traditionally, approaches for defining the transition rules of cellular automata had difficulty to balance the interpretability, accuracy, and convenience. This article presents a new cellular automata model for simulating urban expansion based on random forest algorithm. The proposed model extracts CA transition rules of urban expansion by introducing random factors in training samples and candidate spatial variables that split nodes during the multiple decision trees building process. One significant advantage of our approach is that it can be easily adopted for parallel implementation and has high prediction accuracy and tolerance to random factors in urban expansion. Another strength of the proposed approach is that it can estimate out-of-bag errors to obtain model parameters quickly and measure the importance of spatial variables and explain the contribution of each variable in urban expansion. The model was applied to simulate urban expansion in Foshan City, Guangdong Province. We used the urban land change of 1988 and 2000 as the dependent variable and the spatial variables as the independent variables to construct the CA model based on random forests, then simulate and predict urban expansion of 2000 and 2012. The results show that random forest model can improve the simulation and prediction accuracies by 1.7% and 2.6%, respectively, when compared to the logistic regression model commonly used in CA simulation. This suggests that random forest model is superior for modeling complex nonlinear urban evolution. Urban expansion of Foshan City in 2024 was also predicted according to its urban development trend. Through measuring the importance of some spatial variables that affect urban expansion, we found that distance to national roads and to the city center are the two most important spatial variables for urban expansion simulation in Foshan City.

Key words: random forest, cellular automata, urban expansion, Foshan City

陈凯, 刘凯, 柳林, 朱远辉. 基于随机森林的元胞自动机城市扩展模拟——以佛山市为例[J]. 地理科学进展, 2015, 34(8): 937-946.

CHEN Kai, LIU Kai, LIU Lin, ZHU Yuanhui. Urban expansion simulation by random-forest-based cellular automata: a case study of Foshan City[J]. PROGRESS IN GEOGRAPHY, 2015, 34(8): 937-946.

图/表 9

图1

图2

表1

图3

图4

图5

表2

图6

表3

参考文献 28

1	蔡林霖. 2013. 随机森林的模型选择及其并行化方法[D]. 哈尔滨: 哈尔滨工业大学.
	[Cai L L.2013. Model selection of random forest and its parallelization[D]. Harbin, China: Harbin Institute of Technology.]
2	段晓东, 王存睿, 刘向东. 2012. 元胞自动机理论研究及其仿真应用[M]. 北京: 科学出版社.
	[Duan X D, Wang C R, Liu X D.2012. Cellular automata theory research and simulation applications[M]. Beijing, China: Science Press.]
3	方匡南, 吴见彬, 朱建平, 等. 2011. 随机森林方法研究综述[J]. 统计与信息论坛, 26(3): 32-38.
	[Fang K N, Wu J B, Zhu J P, et al.2011. A review of technologies on random forests[J]. Statistics & Information Forum, 26(3): 32-38.]
4	冯永玖, 刘妙龙, 童小华, 等. 2010. 基于核主成分元胞模型的城市演化重建与预测[J]. 地理学报, 65(6): 665-675.
	[Feng Y J, Liu M L, Tong X H, et al.2010. Kernel principal components analysis based cellular model for restructuring and predicting urban evolution[J]. Acta Geographica Sinica, 65(6): 665-675.]
5	冯永玖, 刘艳, 韩震. 2011. 不同样本方案下遗传元胞自动机的土地利用模拟及景观评价[J]. 应用生态学报, 22(4): 957-963.
	[Feng Y J, Liu Y, Han Z.2011. Land use simulation and landscape assessment by using genetic algorithm based on cellular automata under different sampling schemes[J]. Chinese Journal of Applied Ecology, 22(4): 957-963.]
6	何春阳, 史培军, 陈晋, 等. 2005. 基于系统动力学模型和元胞自动机模型的土地利用情景模型研究[J]. 中国科学: 地球科学, 35(5): 464-473.
	[He C Y, Shi P J, Chen J, et al.2005. Developing land use scenario dynamics model by the integration of system dynamics model and cellular automata model[J]. Science in China: Earth Sciences, 48(11): 1979-1989.]
7	柯新利, 孟芬, 马才学. 2014. 基于粮食安全与经济发展区域差异的土地资源优化配置: 以武汉城市圈为例[J]. 资源科学, 36(8): 1572-1578.
	[Ke X L, Meng F, Ma C X.2014. Optimizing land resource allocation based on food security and regional difference in economic development: a case study in Wuhan metropolitan[J]. Resources Science, 36(8): 1572-1578.]
8	廖江福, 唐立娜, 王翠平, 等. 2014. 城市元胞自动机扩展邻域效应的测量与校准研究[J]. 地理科学进展, 33(12): 1624-1633.
	[Liao J F, Tang L N, Wang C P, et al.2014. Measuring and calibrating extended neighborhood effect of urban cellular automata model based on particle swarm optimization[J]. Progress in Geography, 33(12): 1624-1633.]
9	刘小平, 黎夏, 叶嘉安, 等. 2007. 利用蚁群智能挖掘地理元胞自动机的转换规则[J]. 中国科学: 地球科学, 37(6): 824-834.
	[Liu X P, Li X, Yeh A G O, et al.2007. Discovery of transition rules for geographical cellular automata by using ant colony optimization[J]. Science in China : Earth Sciences, 50(10): 1578-1588.]
10	龙瀛. 2011. 面向空间规划的微观模拟: 数据、模拟与评价[D]. 北京: 清华大学.
	[Long Y.2011. Urban microsimulation for spatial plan: data, modelling, and evaluation[D]. Beijing, China: Tsinghua University.]
11	王云飞, 庞勇, 舒清态. 2013. 基于随机森林算法的橡胶林地上生物量遥感反演研究: 以景洪市为例[J]. 西南林业大学学报, 33(6): 38-45.
	[Wang Y F, Pang Y, Shu Q T.2013. Counter-estimation on aboveground biomass of Hevea brasiliensis plantation by remote sensing with random forest algorithm: a case study of Jinghong[J]. Journal of Southwest Forestry University, 33(6): 38-45.]
12	杨青生. 2008. 地理元胞自动机及空间动态转换规则的获取[J]. 中山大学学报: 自然科学版, 47(4): 122-127.
	[Yang Q S.2008. Dynamic transition rules for geographical cellular automata[J]. Acta Scientiarum Naturalium Universitatis Sunyatseni, 47(4): 122-127.]
13	杨青生, 黎夏. 2007. 基于遗传算法自动获取CA模型的参数: 以东莞市城市发展模拟为例[J]. 地理研究, 26(2): 229-237.
	[Yang Q S, Li X.2007. Calibrating urban cellular automata using genetic algorithms[J]. Geographical Research, 26(2): 229-237.]
14	张鸿辉, 尹长林, 曾永年, 等. 2008. 基于SLEUTH模型的城市增长模拟研究: 以长沙市为例[J]. 遥感技术与应用, 23(6): 618-623.
	[Zhang H H, Yin C L, Zeng Y N, et al.2008. Study on urban growth simulation based on SLEUTH model: Changsha City as an example[J]. Remote Sensing Technology and Application, 23(6): 618-623.]
15	张亦汉, 黎夏, 刘小平, 等. 2013. 耦合遥感观测和元胞自动机的城市扩张模拟[J]. 遥感学报, 17(4): 872-886.
	[Zhang Y H, Li X, Liu X P, et al.2013. Urban expansion simulation by coupling remote sensing observations and cellular automata[J]. Journal of Remote Sensing, 17(4): 872-886.]
16	Breiman L.2001a. Random forests[J]. Machine Learning, 45(1): 5-32.
17	Breiman L.2001b. Statistical modeling: the two cultures[J]. Statistical Science, 16(3): 199-231.
18	Breiman L, Friedman J H, Stone C J, et al.1984. Classification and regression trees[M]. Boca Raton, FL: CRC press.
19	Chen Y M, Li X, Wang S J, et al.2013. Simulating urban form and energy consumption in the Pearl River Delta under different development strategies[J]. Annals of the Association of American Geographers, 103(6): 1567-1585.
20	Genuer R, Poggi J -M, Tuleau-Malot C.2010. Variable selection using random forests[J]. Pattern Recognition Letters, 31(14): 2225-2236.
21	Hastie T, Tibshirani R, Friedman J H.2008. The elements of statistical learning: data mining, inference, and prediction (2nd ed.)[M]. New York, NY: Springer.
22	Kandaswamy K K, Chou K C, Martinetz T, et al.2011. AFP-Pred: a random forest approach for predicting antifreeze proteins from sequence-derived properties[J]. Journal of Theoretical Biology, 270(1): 56-62.
23	Li X, Yeh A G O.2001. Calibration of cellular automata by using neural networks for the simulation of complex urban systems[J]. Environment and Planning A, 33(8): 1445-1462.
24	Li X, Yeh A G O.2002. Neural-network-based cellular automata for simulating multiple land use changes using GIS[J]. International Journal of Geographical Information Science, 16(4): 323-343.
25	Liu X P, Li X, Liu L, et al.2008. A bottom-up approach to discover transition rules of cellular automata using ant intelligence[J]. International Journal of Geographical Information Science, 22(11-12): 1247-1269.
26	Peters J, De Baets B, Verhoest N E C, et al.2007. Random forests as a tool for ecohydrological distribution modelling[J]. Ecological Modelling, 207(2-4): 304-318.
27	Rodriguez-Galiano V F, Ghimire B, Rogan J, et al.2012. An assessment of the effectiveness of a random forest classifier for land-cover classification[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 67: 93-104.
28	Wu F L.2002. Calibration of stochastic cellular automata: the application to rural-urban land conversions[J]. International Journal of Geographical Information Science, 16(8): 795-818.

变量类型	变量	获取方法
因变量	是否转变为城市元胞	遥感分类
空间距离变量	距市中心的距离(x₁)	利用ArcGIS的Distance获取
	距区中心的距离(x₂)
	距镇中心的距离(x₃)
	距高速公路的距离(x₄)
	距省道的距离(x₅)
	距县道的距离(x₆)
自然因素变量	高程栅格图(x₇)
自然因素变量	坡度栅格图(x₈)
限制因素变量	不可发展区域图(x₉)
局部变量	3×3邻域已城市化元胞数(x₁₀)	用ArcGIS的Focal函数

年份		类型	随机森林模型				逻辑回归模型
年份		类型	非城市		城市	精度/%	非城市	城市	精度/%
1988-2000年	遥感分类图	非城市	3444759	220694	94.0		3406151	259302	92.9
	遥感分类图	城市	220008	383507	63.6		253128	350387	58.1
	总精度				89.7				88.0
	Kappa		0.580				0.508
2000-2012年	遥感分类图	非城市	2901659	360751	88.9	2844209	418201	87.2
	遥感分类图	城市	360647	645911	64.2	416346	590212		58.6
	总精度				83.1				80.5
	Kappa		0.531			0.458

[1]	徐明, 石玉立, 王彬. 高分辨率青藏高原历史月降水数据重建[J]. 地理科学进展, 2018, 37(7): 923-932.
[2]	柳林, 刘文娟, 廖薇薇, 余洪杰, 姜超, 林荣平, 纪佳楷, 张政. 基于随机森林和时空核密度方法的不同周期犯罪热点预测对比[J]. 地理科学进展, 2018, 37(6): 761-771.
[3]	戴尔阜, 马良. 土地变化模型方法综述[J]. 地理科学进展, 2018, 37(1): 152-162.
[4]	陈宝芬, 张耀民, 江东. 基于CA-ABM模型的福州城市用地扩张研究[J]. 地理科学进展, 2017, 36(5): 626-634.
[5]	谭敏, 刘凯, 柳林, 朱远辉, 王大山. 基于随机森林模型的珠江三角洲30 m格网人口空间化[J]. 地理科学进展, 2017, 36(10): 1304-1312.
[6]	王海军, 夏畅, 张安琪, 邓羽. 基于约束性CA的大都市郊区城镇增长的情景模拟与管控——以武汉市江夏区为例[J]. 地理科学进展, 2016, 35(7): 793-805.
[7]	顾朝林, 张悦, 翟炜, 管卫华, 李强, 赵娜, 刘晨. 城市与区域定量研究进展[J]. 地理科学进展, 2016, 35(12): 1433-1446.
[8]	刘翠玲, 龙瀛. 京津冀地区城镇空间扩张模拟与分析[J]. 地理科学进展, 2015, 34(2): 217-.
[9]	廖江福, 唐立娜, 王翠平, 许通. 城市元胞自动机扩展邻域效应的测量与校准研究[J]. 地理科学进展, 2014, 33(12): 1624-1633.
[10]	林晨, 周生路, 吴绍华. 基于优化GeoCA模型的土壤侵蚀时空变化模拟——以福建省长汀县为例[J]. 地理科学进展, 2012, 31(7): 911-920.
[11]	龙瀛, 韩昊英, 谷一桢, 沈振江, 毛其智. 城市规划实施的时空动态评价[J]. 地理科学进展, 2011, 30(8): 967-977.
[12]	郭欢欢, 李波, 侯鹰, 孙特生. 元胞自动机和多主体模型在土地利用变化模拟中的应用[J]. 地理科学进展, 2011, 30(11): 1336-1344.
[13]	肖洪1,2\|田怀玉2\|朱佩娟2\|于桓凯3. 基于多智能体的城市人口分布动态模拟与预测[J]. 地理科学进展, 2010, 29(3): 347-354.
[14]	柯新利, 邓祥征, 刘成武. 基于分区异步元胞自动机模型的耕地利用布局优化——以武汉城市圈为例[J]. 地理科学进展, 2010, 29(11): 1442-1450.
[15]	黄宁1, 2\|吝涛3\|章伟婕4\|曹英兰1, 2. 厦门市同安区不同扩展轴上的景观格局梯度分析与比较[J]. 地理科学进展, 2009, 28(5): 767-774.