【摘要】 泥石流灾害是我国最常发生且危害最大的地质灾害之一，因此实现有效、准确的泥石流灾害预测对于人类的生活和生产具有重大意义。研究以四川省石棉县为研究区域，选取12个泥石流影响因子。同时针对传统支持向量机模型精度不高的问题，采用遗传算法、粒子群算法、秃鹰搜索算法以及新型的群智能优化算法—麻雀搜索算法等4种算法来优化支持向量机的超参数C和gamma。通过优化后的支持向量机模型建立泥石流易发性评价模型，同时对比随机森林模型与人工神经网络模型，最后采用受试者工作特征曲线来评价预测模型。研究结果表明，4种优化算法均能有效提高预测准确度，但麻雀搜索算法优化的支持向量机模型具有更高的准确度和受试者工作特征曲线下面积，分别为0.957 3和0.98,并在泥石流易发性分区图中得到验证。因此，麻雀搜索算法优化的支持向量机模型在泥石流易发性评价研究中更为适用。更多还原

【Abstract】 Mudslide hazards are one of the most frequent and hazardous geological hazards in China, so achieving effective and accurate mudslide hazard prediction is of great significance to human life and production. In this study, 12 mudslide impact factors are selected in Shimian county, Sichuan province as the study area. At the same time, four algorithms, including genetic algorithm, particle swarm optimization, bald eagle search algorithm and the new swarm intelligence optimization algorithm-sparrow search algorithm, is used to optimize the hyperparameters C and gamma of the support vector machine. The optimized support vector machine model is used to establish the mudflow susceptibility evaluation model, while comparing the random forest model with the artificial neural network model and finally the subject work characteristic curves to evaluate the prediction model. The results show that all four optimisation algorithms can effectively improve the prediction accuracy, but the support vector machine model optimised by the sparrow search algorithm has higher accuracy and area under the ROC curve of 0.957 3 and 0.98 respectively and is validated in the mudflow susceptibility partition map. Therefore, the support vector machine model optimised by the sparrow search algorithm is more applicable in debris flow susceptibility evaluation studies.更多还原