基于信息熵加权的局部离群点检测算法

SLOM Outlier Mining Algorithm Based on Entropy Weighted

离群点检测是数据挖掘领域的重要研究方向之一,可以从大量数据中发现少量与多数数据有明显区别的数据对象。在诸如网络入侵、无线传感器网络异常事件等检测应用中,离群点检测是一项具有很高应用价值的技术。为了提高离群点检测准确度,文中在局部离群测度(SLOM)算法的基础上,作了一些改进,提出了一种基于密度的局部离群点检测算法ESLOM。引入信息熵确定数据对象的离群属性,并对对象距离采用加权距离,以提高离群点检测准确度。理论分析和实验表明该算法是可行有效的。

Outlier detection is one of the important research field in data mining,which is to find exceptional objects that deviate from the most rest of the data set. And is one of the valuable techniques in many applications, such as network intrusion detection, event detection in wireless sensor network （ WSN ）, and so on. In order to improve the accuracy of detection outliers, made some improvements based on the local outlier measure （ SLOM } algorithm, a density-based local outlier detecting algorithm { ESLOM } is proposed, which educes outlier attributes of each data object by information entropy. Introducing the information entropy determine stray attributes of the data object. And use the weighted distance on the objects distance to improve the outliers detecting accuracy. Theoretical analysis and experimental results show this algorithm is feasible and effective.