基于改进聚类算法的卫星数据异常检测模型

Anomaly Detection Model of Satellite Data Based on Improved Clustering Algorithm

利用数据挖掘技术对Demeter卫星数据进行分析从而发现异常数据已成为当前研究的重点,为了进一步提高异常数据发现的质量,本文提出了一种改进的聚类算法,对Demeter卫星电场数据进行异常检测,该算法首先将数据随机取样,引入信息熵理论,对PAM算法进行改进,并对卫星数据进行划分,以找到聚类中心,最后对改进算法进行了分析与比较,实验结果证明了算法的有效性。

Using data mining technology on Demeter satellite data to analyze and find anomaly data has become the focus of current researching.To further improve the quality of finding anomaly data,the paper proposes a improved clustering algorithm for anomaly detection of demeter satellite electric field data.The algorithm first processes data by random sampling,and introduces information entropy theory,then improving PAM algorithm to divide satellite data,to find clustering Center,Finally,we analyze and compare the improved algorithm.Experiment results show the effectiveness of the algorithm.