Abstract: In this paper，the degree to which numeric data tend to spread is called the dispersion，or variance of the data.It allows us to make better understanding of the data’s overall features，and thus understanding the distribution of data which is used to find the outliers data.A kind of outlier mining methods is improved，based on the deviation of data.The improved algorithm is suitable for multi-dimensional numerical data.

Key words: statistical dispersion, outlier mining, deviation of data

摘要： 对统计数据的散度情况，即数据变异指标，进行了说明，变异指标可以使我们对数据的总体特征有更进一步的了解，进而对数据的分布情况有所了解，变异指标对发现数据中的离群数据有一定的作用。作者使用变异指标对基于偏差的离群数据的发现方法进行改进，改进后的算法适合于多维数值数据。

关键词: 统计变异, 离群数据, 偏差数据