滑坡敏感性的评价结果可以为滑坡防治提供相关判断依据。本文选取重庆市南川区滑坡作为研究对象，对该地区进行滑坡灾害敏感性评价及分区研究，通过影响因子相关性检验、集合模型精度提升和敏感性差异化分析三个方面展开相关研究，取得如下结果：

（1）本文以重庆市南川区地质、水文、气象等多源数据作为基础，共选择高程、坡度、坡向、平面曲率、剖面曲率、泥沙输移指数（STI）、地形湿度指数（TWI）、水流功率指数（SPI）、河流距离、降雨、岩性、断层距离、土壤、土地利用、公路距离和归一化植被指数（NDVI）共计16个影响因子，构建备选数据库。利用确定性系数法（CF）和滑坡点密度分析了南川区滑坡灾害与敏感性因子各分级之间的空间关系。

（2）本文选择皮尔逊相关性检验、多重共线性诊断和相关属性评价三种方法对因子相关性进行检验，结果表明所有的因子均相互独立且对敏感性建模有所贡献。

（3）以误差降低剪枝树（Reduced-error pruning tree，Rept）作为基分类器，分别与Bagging模型、Dagging模型和Real Adaboost模型进行耦合，共得到Rept模型、BRept模型、DRept模型和RRept模型，使用四种模型对重庆市南川区进行了滑坡敏感性评价。使用受试者工作特征曲线（ROC）、标准误差（SE）、95%置信区间（CI）、显著性水平P、平均绝对误差（MAE）对四个模型的精度进行了验证比较，发现所有集合模型的性能优于单一模型，RRept模型在所有比较中展现出了最优的结果。最后通过Wilcoxon符号秩检验的结果表明，最优模型与其他模型都存在显著性差异。

（4）通过kappa系数对所有模型的敏感性图之间的差异程度进行了验证，结果表明，只有Rept模型与BRept模型的敏感性图之间一致性较高。最后使用差异化比较方法分析了最优模型RRept模型与其他模型的敏感性图的差异性。结果表明，RRept模型敏感性图的低估区域主要集中于研究区的西北部，而高估区域主要集中于东南部。同时高估与低估栅格受到影响因子的控制存在一定的空间分布模式。

The evaluation results of landslide susceptibility can provide relevant judgment basis for landslide prevention and control. In this paper, the susceptibility evaluation and zoning study of landslide in Nanchuan district of Chongqing were carried out. In this study, relevant researches were carried out from three aspects: correlation test of influencing factors, accuracy improvement of model and susceptibility differentiation analysis, and the following results were obtained:

(1) Select multi-source data to construct factor alternative database. Based on multi-source geological, hydrological and meteorological data of Nanchuan District, Chongqing, Total choose elevation, slope angle, slope aspect, plane curvature, profile curvature, sediment transport index (STI), topographic wetness index (TWI), Stream Power index (SPI), distance to rivers, rainfall, lithology, distance to fault, soil, landuse, distance to road and normalized difference vegetation index (NDVI) for a total of 16 factors, build alternative database. Based on certainty factor (CF) and landslide point density, the spatial relationship between landslide and influencing factors in Nanchuan district was analyzed.

(2) In this paper, Pearson correlation test, multicollinearity diagnosis and correlation attribute evaluation are selected to test the correlation of factors. The results show that all factors are independent of each other and contribute to susceptibility modeling

(3) The Reduced error pruning tree (Rept) was used as the base classifier, which was coupled with Bagging model, Dagging model and Real Adaboost model respectively. Four integrated models (Rept, BRept, DRept and RRept) were used to evaluate the susceptibility of landslides in Nanchuan District of Chongqing. Receiver operating characteristic curve (ROC), standard error (SE), 95% confidence interval (CI), significance level P and mean absolute error (MAE) were used to verify and compare the accuracy of the four models. The results show that the performance of all integrated models is better than single model. The RRept model showed optimal results in all comparisons. Finally, Wilcoxon signed-rank test shows that there are significant differences between the optimal model and other models.

(4) The degree of difference between susceptibility maps of all models was verified by Kappa coefficient.  The results showed that only Rept model and BRept model had a high consistency between susceptibility maps.  The difference of susceptibility maps between RRept model and other models is analyzed by differentiation comparison method. The results show that the underestimation region of RRept model susceptibility map is mainly concentrated in the northwest of the study area, while the overestimation region is mainly concentrated in the southeast.  At the same time, there is a certain spatial distribution pattern in the influencing factors of overestimation and underestimation grids.

[1]张梁, 张业成, 罗元华. 地质灾害灾情评估理论与实践[M]. 地质出版社, 1998.

[2]殷坤龙,朱良峰.滑坡灾害空间区划及GIS应用研究[J].地学前缘,2001(02):279-284.

[3]许石罗. 基于多源遥感影像的动态滑坡灾害空间预测模型研究[D]. 中国地质大学, 2018.

[4]邱海军. 区域滑坡崩塌地质灾害特征分析及其易发性和危险性评价研究[D].西北大学，2012.

[5]张庭瑜. 府谷县地质灾害易发性分区方法研究[D]. 西安科技大学.

[6] Kreuzer T M, Wilde M, Terhorst B, et al. A landslide inventory system as a base for automated process and risk analyses[J]. Earth Science Informatics, 2017, 10(4): 507-515.

[7] Zhuang J , Peng J , Xu Y , et al. Assessment and mapping of slope stability based on slope units: A case study in Yan'an, China[J]. Journal of Earth System Science, 2016, 125(7):1439-1450.

[8] Chiessi V, Toti S, Vitale V. Landslide susceptibility assessment using conditional analysis and rare events logistics regression: a case-study in the Antrodoco area (Rieti, Italy)[J]. Journal of Geoscience and Environment Protection, 2016, 4(12): 1.

[9] Geographical information systems in assessing natural hazards[M]. Springer Science & Business Media, 2013.

[10]许国庆, 周宇. 基于地貌单元的小区地质灾害易发性分区方法研究[J]. 世界有色金属, 2017 (11): 137-138.

[11]唐川,马国超.基于地貌单元的小区域地质灾害易发性分区方法研究[J].地理科学,2015,35(01):91-98.

[12]刘彬.基于GIS与随机森林算法的斜坡单元类型划分方法[J].经纬天地,2021(04):82-86.

[13]田述军,张珊珊,唐青松,樊晓一,韩培锋.基于不同评价单元的滑坡易发性评价对比研究[J].自然灾害学报,2019,28(06):137-145.

[14]张曦,陈丽霞,徐勇,连志鹏.两种斜坡单元划分方法对滑坡灾害易发性评价的对比研究[J].安全与环境工程,2018,25(01):12-17.

[15]孟田,许晓露,刘汉湖.基于斜坡单元优化的高海拔地区滑坡危险性评价——以金沙江白格滑坡为例[J].河南理工大学学报(自然科学版),2021,40(01):65-73.

[16] Ba Q, Chen Y, Deng S, et al. A comparison of slope units and grid cells as mapping units for landslide susceptibility assessment[J]. Earth Science Informatics, 2018, 11(3): 373-388.

[17] Chen Z, Ye F, Fu W, et al. The influence of DEM spatial resolution on landslide susceptibility mapping in the Baxie River basin, NW China[J]. Natural Hazards, 2020, 101(3): 853-877.

[18]罗金. 基于各类机器学习方法的滑坡易发性评价及软件系统开发[D].长安大学，2021.

[19] Abraham M T, Satyam N, Jain P, et al. Effect of spatial resolution and data splitting on landslide susceptibility mapping using different machine learning algorithms[J]. Geomatics, Natural Hazards and Risk, 2021, 12(1): 3381-3408.

[20] Chang K T, Merghadi A, Yunus A P, et al. Evaluating scale effects of topographic variables in landslide susceptibility models using GIS-based machine learning techniques[J]. Scientific reports, 2019, 9(1): 1-21.

[21]崔阳阳. 基于不同评价单元的滑坡易发性评价方法研究[D].西安科技大学，2021.

[22] Kia M B, Pirasteh S, Pradhan B, et al. An artificial neural network model for flood simulation using GIS: Johor River Basin, Malaysia[J]. Environmental earth sciences, 2012, 67(1): 251-264.

[23] Pradhan B. Flood susceptible mapping and risk area delineation using logistic regression, GIS and remote sensing[J]. Journal of Spatial Hydrology, 2010, 9(2):1-18.

[24]罗杨. 基于GIS与RS技术的什邡市滑坡易发性评价[D].成都理工大学，2019.

[25]郭有金. 基于集成学习算法的西安市滑坡灾害易发性评价[D].西安科技大学，2020.

[26]赵冬梅,角媛梅,邱应美,刘澄静,徐秋娥,张娟.基于maxEnt模型的哈尼梯田核心区滑坡易发性评价[J].水土保持研究,2020,27(04):392-399.

[27] 张向营, 张春山, 孟华君, 等. 基于 GIS 和信息量模型的京张高铁滑坡易发性评价[J]. 地质力学学报, 2018, 24(1): 96-105.

[28] Nhu V H, Mohammadi A, Shahabi H, et al. Landslide detection and susceptibility modeling on cameron highlands (Malaysia): A comparison between random forest, logistic regression and logistic model tree algorithms[J]. Forests, 2020, 11(8): 830.

[29] Huang H P, Yang K C, Lin B W. Statistical evaluation of the effect of earthquake with other related factors on landslide susceptibility: using the watershed area of Shihmen reservoir in Taiwan as a case study[J]. Environmental earth sciences, 2013, 69(7): 2151-2166.

[30]王得双,桂先刚,谭之东,等. 滑坡易发性评价影响因子研究综述[J]. 中国科技论文在线精品论文,2018,11(2):80-91.

[31] Fan W, Wei X, Cao Y, et al. Landslide susceptibility assessment using the certainty factor and analytic hierarchy process[J]. Journal of Mountain Science, 2017, 14(5): 906-925.

[32]李远远,梅红波,任晓杰,胡旭东,李梦迪.基于确定性系数和支持向量机的地质灾害易发性评价[J].地球信息科学学报,2018,20(12):1699-1709.

[33]Wang D， Hao M， Chen S， et al. Assessment of landslide susceptibility and risk factors in China[J]. Natural Hazards， 2021， 108(3): 3045-3059.

[34] Marjanović M, Abolmasov B, Đurić U, et al. Impact of geo-environmental factors on landslide susceptibility using an AHP method: A case study of Fruška Gora Mt., Serbia[J]. Geoloski anali Balkanskoga poluostrva, 2013 (74): 91-100.

[35] Lourembam Chanu M, Oinam B. Impact study for landslide contributing factors using a multi-criterion approach for landslide susceptibility[J]. Arabian Journal of Geosciences, 2021, 14(18): 1-12..

[36] Mind’je R, Li L, Nsengiyumva J B, et al. Landslide susceptibility and influencing factors analysis in Rwanda[J]. Environment, Development and Sustainability, 2020, 22(8): 7985-801.

[37] 范强, 巨能攀, 向喜琼, 等. 基于结果验证的信息量法地质灾害易发性评价——以贵州省开阳县为例[J]. 人民长江, 2015, 46(15): 65-68..

[38] Jebur M N, Pradhan B, Tehrany M S. Optimization of landslide conditioning factors using very high-resolution airborne laser scanning (LiDAR) data at catchment scale[J]. Remote Sensing of Environment, 2014, 152: 150-165.

[39]赵晓萌,蔡新玲,雷向杰,田亮,卫星君.基于Logistic回归的陕南秦巴山区降雨型滑坡预测方法[J].冰川冻土,2019,41(01):175-182.

[40]刘坚,李树林,陈涛.基于优化随机森林模型的滑坡易发性评价[J].武汉大学学报(信息科学版),2018,43(07):1085-1091.

[41]黄发明. 基于 3S 和人工智能的滑坡位移预测与易发性评价[D]. 武汉: 中国地质大学 (武汉)， 2017.

[42]牛瑞卿,彭令,叶润青,武雪玲.基于粗糙集的支持向量机滑坡易发性评价[J].吉林大学学报(地球科学版),2012,42(02):430-439.

[43]唐睿旋,晏鄂川,唐薇.基于粗糙集和BP神经网络的滑坡易发性评价[J].煤田地质与勘探,2017,45(06):129-138.

[44]于宪煜. 基于多源数据和多尺度分析的滑坡易发性评价方法研究[D]. 武汉: 中国地质大学， 2016.

[45]傅贵. 伊犁某典型黄土区滑坡易发性评价研究[D].安徽理工大学，2020.

[46]金帅. 基于3S技术及多模型集成的滑坡易发性评价[D].防灾科技学院，2021.

[47] Brabb E E， Pampeyan E H， Bonilla M G. Landslide susceptibility in San Mateo County， California[R]. US Geological Survey， 1972.

[48]Nilsen T H， Brabb E E. Slope stability studies in the San Francisco Bay region， California[J]. Geol. Soc. Am. Rev. Eng. Geol， 1977， 3: 235-243.

[49] Beaty C B. Landslides and slope exposure[J]. The Journal of Geology， 1956， 64(1): 70-74.

[50]郑玲静,李秀珍,徐瑞池.基于斜坡单元的区域滑坡敏感性评价——以云南省小江流域为例[J].科学技术与工程,2021,21(28):12322-12329.

[51]王家柱,高延超,铁永波,徐伟,白永健,张彦锋.基于斜坡单元的山区城镇滑坡灾害易发性评价：以康定为例[J/OL].沉积与特提斯地质:1-17[2022-06-23].

[52] Chen W， Li W， Chai H， et al. GIS-based landslide susceptibility mapping using analytical hierarchy process (AHP) and certainty factor (CF) models for the Baozhong region of Baoji City， China[J]. Environmental Earth Sciences， 2016， 75(1): 1-14.

[53] Wang Q， Guo Y， Li W， et al. Predictive modeling of landslide hazards in Wen County， northwestern China based on information value， weights-of-evidence， and certainty factor[J]. Geomatics， Natural Hazards and Risk， 2019， 10(1): 820-835.

[54] Zhao C， Chen W， Wang Q， et al. A comparative study of statistical index and certainty factor models in landslide susceptibility mapping: a case study for the Shangzhou District， Shaanxi Province， China[J]. Arabian Journal of Geosciences， 2015， 8(11): 9079-9088.

[55] Xing X， Wu C， Li J， et al. Susceptibility assessment for rainfall-induced landslides using a revised logistic regression method[J]. Natural Hazards， 2021， 106(1): 97-117.

[56]林荣福. 基于优化支持向量机模型的滑坡易发性评价[D].辽宁工程技术大学，2021.

[57] Quan H C， Lee B G. GIS-based landslide susceptibility mapping using analytic hierarchy process and artificial neural network in Jeju (Korea)[J]. KSCE Journal of Civil Engineering， 2012， 16(7): 1258-1266.

[58] Pham B T， Pradhan B， Bui D T， et al. A comparative study of different machine learning methods for landslide susceptibility assessment: A case study of Uttarakhand area (India)[J]. Environmental Modelling & Software， 2016， 84: 240-250.

[59] Phong T V， Phan T T， Prakash I， et al. Landslide susceptibility modeling using different artificial intelligence methods: A case study at Muong Lay district， Vietnam[J]. Geocarto International， 2021， 36(15): 1685-1708.

[60] Wang S , Zhuang J , Zheng J , et al. Application of Bayesian hyperparameter optimized random forest and XGBoost model for landslide susceptibility mapping[J]. Frontiers in Earth Science, 2021, 9: 712240.

[61] Zhu L , Huang L , Fan L , et al. Landslide Susceptibility Prediction Modeling Based on Remote Sensing and a Novel Deep Learning Algorithm of a Cascade-Parallel Recurrent Neural Network[J]. Sensors, 2020, 20(6):1576.

[62]郑迎凯， 陈建国， 王成彬， 等. 确定性系数与随机森林模型在云南芒市滑坡易发性评价中的应用[J]. 地质科技通报， 2020， 39(6): 131-144.

[63]安凯强,牛瑞卿.信息量支持下SVM模型滑坡灾害易发性评价[J].长江科学院院报,2016,33(08):47-51+58.

[64] Pham B T， Prakash I， Singh S K， et al. Landslide susceptibility modeling using Reduced Error Pruning Trees and different ensemble techniques: Hybrid machine learning approaches[J]. Catena， 2019， 175: 203-218.

[65] Hong H， Liu J， Bui D T， et al. Landslide susceptibility mapping using J48 Decision Tree with AdaBoost， Bagging and Rotation Forest ensembles in the Guangchang area (China)[J]. Catena， 2018， 163: 399-413.

[66]杨德宏， 范文. 基于 ArcGIS 的地质灾害易发性分区评价——以旬阳县为例[J]. 中国地质灾害与防治学报， 2015， 26(4): 82-86.

[67]赵艳南， 牛瑞卿. 基于证据权法的滑坡危险性区划探索[J]. 地理與地理信息科學， 2010， 26(6): 19-23.

[68]杨光， 徐佩华， 曹琛， 等. 基于确定性系数组合模型的区域滑坡敏感性评价[J]. 工程地质学报， 2019， 27(5): 1153-1163.

[69] Jaafari A， Najafi A， Pourghasemi H R， et al. GIS-based frequency ratio and index of entropy models for landslide susceptibility assessment in the Caspian forest， northern Iran[J]. International Journal of Environmental Science and Technology， 2014， 11(4): 909-926.

[70] Wosten J H M， van Genuchten M T. Division S-6-Soil and Water Management and Conservation[J]. Using Texture and Other Soil Properties to Predict the Unsaturated Soil Hydraulic Functions， 1988， 52: 1762-1770.

[71] Conforti M， Aucelli P P C， Robustelli G， et al. Geomorphology and GIS analysis for mapping gully erosion susceptibility in the Turbolo stream catchment (Northern Calabria， Italy)[J]. Natural hazards， 2011， 56(3): 881-898.

[72] Moore I D， Grayson R B， Ladson A R. Digital terrain modelling: a review of hydrological， geomorphological， and biological applications[J]. Hydrological processes， 1991， 5(1): 3-30.

[73]江青龙， 谢永生， 张应龙， 等. 基于 GIS 与 RS 小流域空间数据挖掘[J]. 水土保持研究， 2010， 17(6): 64-67.

[74] Zhang Y， Wu W， Qin Y， et al. Mapping landslide hazard risk using random forest algorithm in Guixi， Jiangxi， China[J]. ISPRS International Journal of Geo-Information， 2020， 9(11): 695.

[75]罗渝. 降雨滑坡形成机理及预测方法研究[D]. 北京. 中国科学院研究生院,2013.

[76]邵江， 许吉亮. 一种断层影响基岩滑坡的失稳机理和稳定性分析[D]. ， 2008.

[77] Zhang G， Cai Y， Zheng Z， et al. Integration of the statistical index method and the analytic hierarchy process technique for the assessment of landslide susceptibility in Huizhou， China[J]. Catena， 2016， 142: 233-244.

[78]陈朝亮. 基于 GIS 的内江市地质灾害易发性评价研究[D]. 西南科技大学， 2019.

[79] Shortliffe E H， Buchanan B G. A model of inexact reasoning in medicine[J]. Mathematical biosciences， 1975， 23(3-4): 351-379.

[80] Heckerman D. Probabilistic interpretations for MYCIN's certainty factors[M]//Machine intelligence and pattern recognition. North-Holland， 1986， 4: 167-196.

[81]尚敏， 廖芬， 马锐， 等. 白家包滑坡变形与库水位， 降雨相关性定量化分析研究[J]. 工程地质学报， 2021， 29(3): 742-750.

[82] Witten I H, Frank E, Hall M A, et al. Practical machine learning tools and techniques[C]//Data Mining. 2005, 2(4).

[83] Quinlan J R. Simplifying decision trees[J]. International journal of man-machine studies， 1987， 27(3): 221-234.

[84] Pham B T， Prakash I. A novel hybrid intelligent approach of random subspace ensemble and reduced error pruning trees for landslide susceptibility modeling: A case study at mu cang chai district， yen bai province， viet nam[C]//International Conference on Geo-Spatial Technologies and Earth Resources. Springer， Cham， 2017: 255-269.

[85]Khosravi K， Pham B T， Chapi K， et al. A comparative assessment of decision trees algorithms for flash flood susceptibility modeling at Haraz watershed， northern Iran[J]. Science of the Total Environment， 2018， 627: 744-755.

[86] Breiman L. Bagging predictors[J]. Machine learning， 1996， 24(2): 123-140.

[87] Tien Bui D， Pradhan B， Revhaug I， et al. A comparative assessment between the application of fuzzy unordered rules induction algorithm and J48 decision tree models in spatial prediction of shallow landslides at Lang Son City， Vietnam[M]//Remote sensing applications in environmental research. Springer， Cham， 2014: 87-111.

[88] Chen W, Shahabi H, Zhang S, et al. Landslide susceptibility modeling based on gis and novel bagging-based kernel logistic regression[J]. Applied Sciences, 2018, 8(12): 2540.

[89] Kotsianti S B, Kanellopoulos D. Combining bagging, boosting and dagging for classification problems[C]//International Conference on Knowledge-Based and Intelligent Information and Engineering Systems. Springer, Berlin, Heidelberg, 2007: 493-500.

[90] Kai M T , Witten I H . Stacking bagged and dagged models[C]// International Conference on Machine Learning. Morgan Kaufmann Publishers Inc. 1997.

[91]Li J R， Huang T. Predicting and analyzing early wake-up associated gene expressions by integrating GWAS and eQTL studies[J]. Biochimica et Biophysica Acta (BBA)-Molecular Basis of Disease， 2018， 1864(6): 2241-2246.

[92]Freund Y， Schapire R E. A decision-theoretic generalization of on-line learning and an application to boosting[J]. Journal of computer and system sciences， 1997， 55(1): 119-139.

[93]Micheletti N， Foresti L， Robert S， et al. Machine learning feature selection methods for landslide susceptibility mapping[J]. Mathematical geosciences， 2014， 46(1): 33-57.

[94]Schapire R E， Singer Y. Improved boosting algorithms using confidence-rated predictions[J]. Machine learning， 1999， 37(3): 297-336.

[95]Ko F W Y， Lo F L C. From landslide susceptibility to landslide frequency: A territory-wide study in Hong Kong[J]. Engineering geology， 2018， 242: 12-22.

[96] Chen W, Li Y, Tsangaratos P, et al. Groundwater spring potential mapping using artificial intelligence approach based on kernel logistic regression, random forest, and alternating decision tree models[J]. Applied Sciences, 2020, 10(2): 425.

[97]Hong H， Liu J， Zhu A. Landslide susceptibility evaluating using artificial intelligence method in the Youfang district (China)[J]. Environmental Earth Sciences， 2019， 78(15): 1-20.

[98] Wang G, Chen X, Chen W. Spatial prediction of landslide susceptibility based on GIS and discriminant functions[J]. ISPRS International Journal of Geo-Information, 2020, 9(3): 144.

[99]Tien Bui D， Tuan T A， Klempe H， et al. Spatial prediction models for shallow landslide hazards: a comparative assessment of the efficacy of support vector machines， artificial neural networks， kernel logistic regression， and logistic model tree[J]. Landslides， 2016， 13(2): 361-378.

[100]Tien Bui D， Ho T C， Pradhan B， et al. GIS-based modeling of rainfall-induced landslides using data mining-based functional trees classifier with AdaBoost， Bagging， and MultiBoost ensemble frameworks[J]. Environmental Earth Sciences， 2016， 75(14): 1-22.

[101]Xiao T， Segoni S， Chen L， et al. A step beyond landslide susceptibility maps: a simple method to investigate and explain the different outcomes obtained by different approaches[J]. Landslides， 2020， 17(3): 627-640.