基于OCSVM-RF耦合模型的雅江县地质灾害易发性评价

Evaluation of geological disaster susceptibility in Yajiang County based on OCSVM-RF model

地质灾害威胁着工农业安全生产及人民生命财产安全,因此,对地质灾害易发性进行高效率、高精度的评价尤为重要。以四川省雅江县为研究区,机器学习方法为基础,分别构建了传统逻辑回归模型、随机森林模型和OCSVM-RF耦合模型,对雅江县地质灾害易发性进行评价并制作雅江县地质灾害易发性分级图。同时使用ROC和AUC值对评价精度进行分析,得到AUC值分别为:逻辑回归(0.9017),随机森林(0.9523),OCSVM-RF耦合模型(0.9575),得出OCSVM-RF耦合模型的评价精度明显高于传统机器学习模型。因此,可以使用该耦合模型对地质灾害易发性进行评价,为研究区的防灾减灾工作提供一定的理论参考。

Geological disasters pose a threat to the safety of industrial and agricultural production,as well as the safety of people’s lives and property.Therefore,it is particularly important to conduct efficient and high-precision evaluations of the susceptibility of geological disasters.Based on the machine learning method of Yajiang County in Sichuan Provience as the study area,the traditional logical regression model,random forest model and OCSVM-RF coupling model were constructed respectively to evaluate the vulnerability of geological disasters in Yajiang County and make a classfication map of the vulnerability of geological disasters in Yajiang County.At the same time,ROC and AUC values were used to analyze the evaluation accuracy,and AUC values obtains as follows:logical regression(0.9017),random forest(0.9523),OCSVM-RF(0.9575).It was conclued that the evaluation accuracy of OCSVM-RF coupling model was significantly higher than that of traditional machine learning model.Therefore,the coupling model can be used to evaluate the susceptibility of geological disasters,providing a certain theoretical reference for disaster prevention and reduction work in the research area.