铁路货车静载重发展变化的数据挖掘研究  

Data Mining on Index of Static Load of Freight Cars

作　　者：刘智丽[1,2] 周熙霖[3] 方琪根[3] 

机构地区：[1]北京交通大学城市交通复杂系统理论与技术教育部重点实验室,北京100044 [2]北京交通大学中国综合交通研究中心,北京100044 [3]中铁第四勘察设计院集团有限公司,武汉430063

出　　处：《交通运输系统工程与信息》2012年第4期128-134,共7页Journal of Transportation Systems Engineering and Information Technology

基　　金：国家自然科学基金(71131001);中央高校基本科研业务费专项资金资助(2011JBM251);青年骨干教师出国研修资助(2010709521)

摘　　要：铁路货车静载重是衡量货车运用效率的一项重要指标,在线路设计中也是预测静载重系数的重要依据.在对货车静载重指标数据结构进行梳理的基础上,采用K-medorids聚类算法进行数据挖掘,发现铁路局静载重水平具有明显的分级特征,而且各个铁路局在静载重分级中的位置相对稳定.通过因素分析发现静载重不同的货物品类发送比例的差异是引起局间静载重差异的主要影响因素.数据挖掘结果表明,在进行车流折算时区域间的静载重差异不应忽视,宜采用分区域选取静载重系数的方法.该研究成果对于指导铁路货车静载重系数预测具有重要意义.Static load of freight vehicles is one of the important indices to evaluate the efficiency of railway wagon usage, as well an important basis for predicting coefficient of static load in railway line design. Based on sorting out the data structure of static load, clustering algorithm of K-medorids is empLoyed for data mining to lind the significant and relatively stable classification characteristics between static load ot＂ raway administrations. It is diftrent delivery proportions of heavy goods that are indicated by index factor analysis to cause the static load classification of railway administrations. The data-mining also indicates the differences of static load between difrent grading railway administrations which cannot be ignored, theretbre determining the coefficient of static load in railway design based on divided region is recommended.

关 键 词：铁路运输 货车静载重 聚类算法 指标因素分析 

分 类 号：U268.6[机械工程—车辆工程]