摘要： 在给数据挖掘这类应用准备数据的过程中，面临着一系列数据清洗问题。要把数 据清洗过程做得很灵活并不容易，已有的工具往往过于依赖特定的应用。该文提出并实现 了 一个可扩展的数据清洗框架。它以术语模型、过程描述文件、共享库等概念和技术实现 了模 块的高度独立性和系统的可扩展性，并提供了一个可视化的流程定义环境。

关键词: 数据准备, 数据清洗, 术语模型

Abstract: There exist many data cleaning tasks on preparing data for applicati ons like data mining. It's not trivial to make data cleaning process flexible e n ough. This paper presents an open and extensible framework for data cleaning. It gains its extensibility by employing innovative features like term model, pr oce ss description file and rule&Dic repository. The system has also implemented a v isual environment to define data cleaning flow.

中图分类号: 

• TP311.12