高维空间数据的一种异常检测方法研究

A Algorithm of Outier Detection for High Dimensional Data

就布尔型的高维数据提出了一种异常检测方法,通过定义反映数据稀疏程度的覆盖系数,搜索其低维子空间里的异常模式来检测异常。该方法使用遗传算法来优化搜索过程,并取得了较好的结果。

In this paper, we propose a algorithm of outlier detection for high dimensional data whose attributes are all boolean. It searches outlier mode in low dimensional subspace by defining a covered coefficient which can reflect the sparse degree of data. In this algorithm, we optimize our search process by using genetic algorithm and get a good iesult.