摘要： 孤立点检测是知识发现中的一个活跃领域，如信用卡欺诈、入侵检测等。研究孤立点的异常行为能发现隐藏在数据集中更有价值的知识。该文提出基于局部偏离因子(LDF)的孤立点检测算法，利用每个数据点的LDF衡量该数据点的偏离程度。实验结果表明，该算法能有效检测孤立点，其效率高于LSC算法。

关键词: 孤立点, k距离邻居, 局部偏离因子

Abstract: Outlier detection is a hot research field in knowledge discovery in databases, such as credit card fraud, and intrusion detection, etc. Finding the rare abnormal behaviors or the outliers can be more interesting than finding the common patterns. This paper proposes a new outlier detection algorithm based on Local Deviation Factor(LDF). This algorithm counts the number of each point’s LDF to reflect its isolation degree. The experimental results show that this algorithm can efficiently detect outliers and has higher efficiency than outliers detection algorithm LSC.

Key words: outlier, k-distance neighbors, local deviation factor

中图分类号: 

• TP311