Data Mining And Application Research In Railway Marshalling Yard

Railway marshalling yard is a station which in railway networksconcentrates on dealing with freight train’s arrival, disintegration,marshalling departure, direct line and other operations, and for theseoperations with a relatively perfect shunting function. With thedevelopment of computer technology and the implementation of railwaytransportation management information system (TMIS), the amount ofbusiness data is increasingly larger, and valuable relationships and rulesare hiding in these mass data. Analyzing these data isn’t only for the needof research, more importantly is to provide intelligent decision supportinformation to railway management departments for railway marketingand transport safety. This paper introduces mining technologies tocustomer segmentation system and security risk early warning system ofrailway marshalling yard, and the main research work is as follows:First，this paper proposes an improved K-means algorithm. Thetraditional K-means algorithm that identifies isolated points with distancecannot do with density non-uniform data sets local features. In view ofthis disadvantage we improve the algorithm, and propose an isolatedpoints clustering and the introduction of variance is processed on theisolated point. It is verified by experiments that the improved K-meansalgorithm can effectively avoid the blindness of abandoning all theisolated points, can effectively reduce the total squared error, and thatdesignating data of the same trait into a cluster can improve the clusteringaccuracy. Applying the improved K-means algorithm to railway customersegmentation makes clustering results more accurate, and from amultidimensional perspective comprehensively and thoroughlysubdivides customer consumption behavior.Second, Apriori algorithm is optimized. The algorithm efficiency isimproved by reducing frequency of scanning database and decreasing thenumber of candidate sets. Through illegal licensing historical data mining,the association rule of data items such as workshop, post and shift isfound out and the regularity of field worker violations is summed up, thereby the scientific foundation of making proper prevention andmanagement decisions is provided to railway transportation safetymanagement departments.