论文部分内容阅读
区域泥石流危险度评价是环境保护和减灾的重要内容。针对泥石流危险度评价的多因素影响问题,通过全面收集有代表性的泥石流样本资料,应用免疫进化算法,建立了区域泥石流危险度评价的投影寻踪聚类模型。研究结果表明,该模型依据样本自身的数据特性寻求最优的投影方向,并通过投影方向计算反映评价样本综合特征信息的投影特征指标。该模型不但可以确定各评价指标的权重,避免权重人为的任意性,还能在区域泥石流不同危险度级别划分的基础上,以平均贡献率的大小揭示不同影响因子对评价结果的影响,从而可清楚地筛选出其中的关键影响因子。此研究从定量的角度进一步深化了对泥石流危险度评价影响因子的认识。
Regional debris flow hazard assessment is an important part of environmental protection and disaster reduction. In order to solve the multi-factor impact of debris flow hazard assessment, a projection pursuit complex clustering model for regional debris flow hazard assessment was established by collecting data of typical debris flow samples and applying immune evolutionary algorithm. The results show that this model seeks the optimal projection direction according to the data characteristics of the sample itself and calculates the projection characteristic index which reflects the comprehensive feature information of the evaluation sample through the projection direction. The model can not only determine the weight of each evaluation index, avoid the arbitrariness of weight, but also reveal the influence of different impact factors on the evaluation results based on the division of different risk levels of debris flow in the area, Clearly identify the key impact factors. This study further deepens the understanding of the influencing factors of debris flow risk assessment from a quantitative point of view.