【摘要】 提出一种基于单维分割的高维数据聚类算法HDCA-SDP,该算法利用单维空间能划分数据的性质,对整个数据集进行逐维聚类,解决了传统聚类算法带来的维度困扰问题,对数据集大小和数据空间维数具有良好的可伸缩性,且聚类结果的精度比传统的高维聚类算法有较大的提高.实验结果表明,该算法在处理高维大规模数据时是有效的.更多还原

【Abstract】 A new algorithm(HDCA-SDP) based on single dimension partition is proposed.This algorithm makes use of character of data partition in single dimension and clustering on the whole dataset by each dimension.The problem of dimensionality curse can be handled well.It has good scalability on the size and dimensionality of data and more precision than traditional algorithms.Experiment results show that HDCA-SDP is efficiency when clustering on high dimensional data.更多还原