基于改进PPC模型的铁路弃渣场综合风险评价系统

为了客观准确地评估铁路弃渣场的综合风险，建立基于改进投影寻踪聚类（PPC）模型的风险评价模型。首先建立了6个维度共19个因素的指标体系，并制定相应评价标准；然后借鉴K均值聚类思想来确定PPC模型关键系数-密度窗宽R，以解决传统方法造成的聚类效果差等问题；同时将遗传算法（GA）作为模型优化算法得到最优投影方向和样本投影值等关键数据；最后以贵昆线铁路的10座弃渣场作为案例研究，研究结果表明:“弃渣场边坡条件”和“地形与地基条件”2个维度对弃渣场的综合风险影响最大；验证了所建模型的科学性，评价结果能够规避人为主观因素的干扰，更加与实际情况相贴合。

Comprehensive risk assessment system for railway abandoned dreg site based on improved PPC model

In order to objectively and accurately evaluate the comprehensive risk of railway abandoned dreg site,a risk assessment model based on the improved projection pursuit clustering (PPC) model was established.Firstly,an index system involving 19 factors in 6 dimensions was established,and the corresponding evaluation criteria were formulated.Then,the key coefficient of PPC model,the window radius R of local density,was determined by using the K means clustering method,so as to solve the problems such as poor clustering effect caused by traditional methods.At the same time,the genetic algorithm (GA) was used as the model optimization algorithm to obtain the key data such as the optimum projection direction and sample projection value.Finally,ten abandoned dreg sites of Guizhou Kunming railway were used as the case studies.The results showed that the two dimensions of “slope condition” and “topography and foundation condition” had the greatest influence on the comprehensive risk of abandoned dreg site.The scientific nature of the established model was verified.The evaluation results could avoid the interference of human subjective factors and were more in line with the actual situation.