基于GIS和PSO-SVM模型的文山州石漠化风险评估  

Risk Assessment of Rocky Desertification in Wenshan Prefecture Based on GIS and PSO-SVM Model

作　　者：冷信风 赖祖龙[2] 

机构地区：[1]修水县国土资源局,江西修水332400 [2]中国地质大学信息工程学院,湖北武汉430074

出　　处：《安全与环境工程》2014年第4期19-24,共6页Safety and Environmental Engineering

基　　金：国家高技术研究发展计划(863计划)项目(2007AA092102)

摘　　要：针对喀斯特石漠化风险评估影响因素较复杂且涉及众多地理空间对象的特点,本文在采用GIS技术对空间数据进行管理分析的基础上,提出了基于粒子群优化的支持向量机(PSO-SVM)的石漠化风险评估模型,分析了基于GIS的石漠化风险评估指标体系的建立过程,详细探讨了主成分分析与支持向量机模型相结合进行石漠化风险评估的方法与实现过程,并以云南省文山州为研究实例,将其与单独SVM模型和PSO-BP模型风险评估结果进行了对比分析,结果表明PSO-SVM模型的石漠化风险评估精度较高,效果更好。The influence factors of risk assessment for Karst rocky desertification are complex and involve many geographical spatial objects.This paper applies GIS technology to managing spatial data,puts forword supporting vector machine model based on particle swarm optimization（PSO-SVM）for rocky desertification risk assessment,analyzes the establishment of the rocky desertification risk assessment index system based on GIS,and discusses risk assessment method for rocky desertification and implementation process in detail based on the combination of principal component analysis and support vector machine model.At the end,the paper takes Wenshan Prefecture in Yunnan Province as a case study,and compares the results of PSO-SVM model with the results of single SVM model and PSO-BP model,whick shows PSO-SVM model for rocky desertification risk assessment is more accurate and effective.

关 键 词：喀斯特石漠化 风险评估 粒子群优化(PSO)算法 支持向量机(SVM) GIS 

分 类 号：X43[环境科学与工程—灾害防治] X144