Abstract: There are rich information resources in the railway freight invoice. It is important and basic data for the prosecution and management of the railway production. This paper from the three aspects knowledge excavates the freight invoice. The characteristic relative model is developed and the characteristic rule knowledge expression is abstracted by applying set theory. It presents the algorithms, knowledge excavates the practical freight invoice and analyzes the inspiration of marketing; applies the clustering to knowledge excavate the model and does research for the subdivision of goods transportation market from such aspects as the transportation distance, the sort of goods, the number of price for goods transportation and the clients’ incoming contributions to the corporation etc. The knowledge from excavating has specified the choice of object markets for the railway goods transportation. The freight invoice data is handled by ARIMA method to be seasonal knowledge excavated. The freight amount in 2006 is evaluated according to the historical data from 1999 to 2005.

Key words: freight invoice, knowledge discovery, clustering analysis, subdivision of market

摘要： 铁路货票蕴含着极为丰富的信息资源，它是铁路生产经营管理的重要基础数据。该文从3个方面对货票数据进行了知识挖掘。运用集合理论构造关系数据库特征关联模型，描述了特征规则知识的表达，提出了算法，并对实际的货票数据进行了知识挖掘，分析了知识对营销的启示；运用聚类知识挖掘模型，从货物运距、货物运价号和客户对公司的收入贡献等方面探讨了货运市场的细分，挖掘出来的知识明确了铁路货运目标市场的选择；运用ARIMA模型对货票数据进行了季节性知识挖掘，用1999年—2005年的历史数据估算2006年的货运量。

关键词: 货票数据, 知识发现, 聚类分析, 市场细分

CLC Number: 

• TP18