摘要： 描述了信用卡数据分类挖掘中的数据预处理过程,先从纷杂的应用数据中选择 与分析主题有关的数据表,经过一系列数据库操作,从这些数据表中获得样本的输入信息, 然后从交易历史信息中抽象出对样本分类的相关参数,将分类标准量化后得到对样本的分类 ?最后经过数据规格化?异常处理?缺失处理等数据清理方法对样本数据进行加工,得到分 类挖掘所需的样本数据?

关键词: 信用卡, 分类, 挖掘, 数据预处理

Abstract: This paper describes the data preprocessing of classification minin g of credit card. The database tables concerned with analysis topic from many ap plication data are selected, and the input information of training samples from these tables through a set of database operations can be got. Then, many parame ters used to measure classification standard from history transaction data are a bstracted, and training samples are classified manually. At last, the paper hand les with problems in the training data such as noisy data, missing values and no rmalization, and obtains training sample data needed by classification mining fi nally.

中图分类号: 

• TP399