论文部分内容阅读
为了实现斜坡地质灾害易发性评价能够随着影响因子动态变化和灾害点数目不断增加的自适应更新,使得评价精度能够随着影响因子认识的深入和样本点的丰富不断升高,提出一种数据驱动的斜坡地质灾害易发性评价系统自适应更新机制,以影响因子处理与筛选、支持向量机以及正确率、Kappa系数和AUC三个评价指标为理论支撑,设计并实现斜坡地质灾害易发性评价系统。该系统由数据预处理、易发性评价模型构建、评价模型评估、地质灾害易发性评价等4个模块构成,以数据变化触发系统自适应更新,可以支持影响因子的动态更新、地质灾害样本点的实时更新以及评价模型的迭代更新,使得评价模型能动态升级以适应最新的区域环境特征和地质灾害特征,保证评价模型的准确性。在浙江省文成县的实验结果表明,数据驱动自适应更新的斜坡地质灾害易发性评价系统能对地质灾害进行有效评价,并随着数据的丰富,将有效提高评价精度。
In order to realize the adaptive evaluation of slope geohazard vulnerability with the dynamic change of influence factors and the increasing number of disaster points, the evaluation accuracy can be improved with the understanding of influence factors and the increase of sample points The data-driven slope geohazard vulnerability assessment system adaptive update mechanism is based on the theory of influencing factor processing and screening, support vector machines, correctness, Kappa coefficient and AUC. The design and implementation of slope geological disasters Sexual evaluation system. The system consists of four modules: data preprocessing, vulnerability assessment model building, evaluation model evaluation and geohazard vulnerability assessment. Data update triggers the system to be adaptively updated to support dynamic update of impact factors, geological disaster samples The real-time update of points and the iterative updating of evaluation models make the evaluation models dynamically upgrade to adapt to the latest regional environmental features and geological hazard characteristics and ensure the accuracy of evaluation models. The experimental results in Wencheng County of Zhejiang Province show that the data-driven adaptively updated slope geohazard vulnerability assessment system can effectively evaluate the geological disasters and improve the evaluation accuracy with the enrichment of data.