Abstract: The dispose of the noisy data is an important part in data cleaning. This paper presents a clustering-based method to dispose noisy data on the level of attribute. This method finds the attribute that makes noise by clustering analysis. Meanwhile, it gets the noise trend in the attributes and this information can be used subsequently to improve the quality of the data.

摘要： 噪声数据的处理是数据清洗的一个重要环节。文章提出一个基于聚类分析的在属性级别上处理噪声数据的算法。该算法通过聚类分析确认噪声数据产生噪声的具体属性,同时统计噪声在属性上的分布规律,这些统计数据可以在下一步的工作中进一步提高数据质量。

关键词: 数据清洗, 聚类, 噪声数据, 数据质量

CLC Number: 

• TP311.13