关于航空器异常数据检测仿真研究  

Aircraft Abnormal Data Mining Method Based on Fuzzy Closer Algorithm

作　　者：姜滨[1] 杨杰明[2] 

机构地区：[1]北华大学师范分院,吉林吉林132000 [2]东北电力大学信息工程学院,吉林吉林132012

出　　处：《计算机仿真》2015年第12期72-75,共4页Computer Simulation

摘　　要：在对航空器的异常数据检测中,由于数据间特征相似性和多样性,使得数据之间的内在关联性高。采用传统的检测方法时,由于这种高的关联性对检测数据的控制参数产生干扰,导致挖掘效率低,误差大的问题,提出一种基于改进模糊贴近算法的航空器异常数据挖掘方法,分析了航空器异常数据挖掘原理,结合模糊数学的聚类概念,确定检测数据初始聚类中心和控制参数,通过迭代公式完成聚类分析,引入贴近指标权向量,依据指标权向量对聚类准则函数进行塑造,完成异常数据聚类中心的分解及合并,并应用于航空器异常数据挖掘中,实验结果表明,引入模糊贴近规则后,提高了航空器异常数据挖掘的可靠性。In the abnormal data detection for aircraft, the similarity and diversity between data characteristics make the high internal connection between the data. Traditional detection methods, owing to the high correlation of interference on the testing data of the control parameters, lead to low mining efficiency and large error. An aircraft abnormal data mining method is presented based on the improved fuzzy closer algorithm, the aircraft abnormal data mining theory is analyzed, combined with the concept of fuzzy clustering, the initial clustering center and the control parameters are determined, and the cluster is analyzed through the iterative formula. A closer index weight vector is introduced, a clustering criterion function is obtained according to the index weight vector, the clustering center of abnormal data is decomposed and merged, and the mining of aircraft abnormal data is comleted. Experimental results show that the proposed method can improve the reliability of the aircraft data mining.

关 键 词：模糊贴近 航空器 异常数据挖掘 

分 类 号：F127[经济管理—世界经济]