为了提高高维数据聚类精度,提出了一种基于数据分布规律的K-means聚类方法。通过K-means聚类粗略寻找高维数据分布规律,构造不同的自适应因子对聚类数据进行综合K-means聚类精度校正。将所提出方法应用于平台惯导系统标定数据聚类中,计算结果表明该方法可以很好的对加速度计标定数据进行聚类和评价,具有较好的实际应用价值。

To improve the precision of high-dimensional data cluster, an improved K-means clustering algorithm based on data distribution was proposed. The distribution of high-dimensional data was found by the K-means clustering method. The clustering accuracy of the data was corrected by the clustering factor. The proposed method is applied in the platform inertial navigation system (INS); the results show that the method is good for calibration data clustering and evaluation.