基于平衡核函数聚类的飞行航迹数据分析方法

Flight Track Data Analysis Method Based on Balanced Kernel Function Clustering

摘要:

目前,使用数据挖掘的方法对目标的飞行航迹进行分析来确定航迹类别具有许多应用价值。飞行航迹数据具有维数高、交连多、可分类性能差等特点,要做到尽可能精确的聚类和分类十分困难。文章立足提高飞行航迹数据聚类分析的准确性,在航迹特征数据的预处理阶段,提出了一种平衡核函数的 K-均值聚类方法,可以解决高维特征数据带来的奇异性,还能提高交叠样本的聚类性能；设计了一种模糊支持向量机的算法框架实现航迹的分类。通过实际飞行航迹数据集测试了设计框架下航迹聚类和分类识别的有效性,在实际工程上具有广泛的应用前景。

Abstract:

At present, using data mining method to analyze the flight track to determine the track type has many applica?tion values. The flight track data has the characteristics of high dimension, multi-link and poor classification performance,so it is very difficult to cluster and classify as accurately as possible. In this paper, aiming to improve the accuracy of flighttrack data clustering analysis, in the preprocessing stage of track characteristic data, a K-MEANS clustering method basedon balanced kernel function was proposed. It can solve the singularity of high dimensional feature data and improve theclustering performance of overlapping samples. A fuzzy and support vector machine algorithm framework was designed torealize track classification. The validity of track clustering and classification was tested by the actual flight track data set,which had a wide application prospect in practical engineering.