基于随机森林模型的重庆市滑坡灾害的研究

发布时间：2018-06-02 01:38

本文选题：滑坡灾害 + 重庆市　；参考：《重庆师范大学》2015年硕士论文

【摘要】：滑坡属于一种高发并且带来严重危害的地质灾害,滑坡带来的危害包括巨额的经济损失,以及惨痛的人员伤亡,并且影响社会的安定。我国滑坡灾害分布广泛且发生频繁,尤其是处于我国第二阶梯前缘的重庆,区域内地形支离破粹,地质结构复杂,降雨量充沛。随着西部大开发战略的实施以及三峡库区的建立,人类活动随之变得越来越频繁,使得区域内滑坡的灾害发生也越来越频繁。因此,为了给滑坡灾害防治提供有力的依据,研究区域内滑坡灾害具有十分重要的理论意义和现实意义。具体研究内容如下:(1)本文简单介绍了目前滑坡灾害造成的危害,以及研究滑坡的意义,国内外对于滑坡研究的进展和现状;(2)基于重庆市2000年到2010年的滑坡灾害发生历史数据,通过分析重庆市地形地貌即坡度、坡向、高程,以及降雨量,其中降雨量选取2000年到2010年的月降雨量,简要阐述这些指标对于滑坡发生所造成的影响;(3)通过对比决策树算法、bagging和SVM算法之间的优缺点,发现随机森林算法随着数据规模的增大以及数据属性纬度的增加,能够快速、高效的完成分类预测。因此,本文最终选取随机森林算法建立滑坡预测模型,尝试能够更加准确的进行滑坡研究与预测。其中随机森林建模过程以R语言为实现工具。研究结果表明,模型训练样本及测试样本的准确率均在80%以上,同时模型对于特征的重要性评估研究发现,坡度对于滑坡灾害的发生影响最大,坡度在15度到30度之间极易发生滑坡,其中降水也是诱发滑坡灾害不可忽视的重要因素,受强降雨季节性活动的影响,滑坡具有明显的季节性,多发生在每年的7月到10月。
[Abstract]:Landslide is a kind of geological disaster with high incidence and serious harm. The hazards brought by landslide include huge economic losses, heavy casualties, and influence on social stability. Landslide disasters are widely distributed and occur frequently in China, especially in Chongqing, which is in the front of the second step of our country. The topography of the region is broken apart, the geological structure is complex, and the rainfall is abundant. With the implementation of the western development strategy and the establishment of the three Gorges Reservoir area, human activities become more and more frequent, which makes the landslide disasters in the region more and more frequent. Therefore, in order to provide a strong basis for landslide prevention and control, it is of great theoretical and practical significance to study landslide disasters in the region. The specific research contents are as follows: (1) this paper briefly introduces the hazards caused by the landslide disaster at present, and the significance of the research on landslide, the progress and present situation of landslide research both at home and abroad) based on the historical data of landslide disasters occurring from 2000 to 2010 in Chongqing. By analyzing the topography and landform of Chongqing, that is, slope, slope direction, elevation, and rainfall, the monthly rainfall from 2000 to 2010 is selected. By comparing the advantages and disadvantages between the decision tree algorithm and the SVM algorithm, it is found that the stochastic forest algorithm can quickly increase the size of the data and the latitude of the data attributes. Efficient implementation of classification prediction. Therefore, in this paper, random forest algorithm is selected to establish landslide prediction model, which can be more accurate for landslide research and prediction. In the process of stochastic forest modeling, R language is used as the implement tool. The results show that the accuracy of model training samples and test samples are above 80%, and the evaluation of the importance of the model to the characteristics shows that the slope has the greatest influence on the occurrence of landslide disaster. The slope between 15 degrees and 30 degrees is prone to landslide, among which precipitation is also an important factor that can not be ignored. Due to the seasonal activities of heavy rainfall, landslides have obvious seasonality, most of which occur from July to October of each year.
【学位授予单位】：重庆师范大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P642.22

【相似文献】