摘要： 高维稀疏数据的聚类分析是目前数据挖掘领域内亟待解决的问题之一。传统的聚类方法中，大部分不适用于高维稀疏数据，不能得到满意的结果。该文借助对象组相似度和对象组的特征向量，提出了一种实现聚类的方法。根据聚类结果后，根据聚类集合的上确界和下确界给出新对象的分类。该方法思想明了，实现起来简单轻松，结果准确可靠。

关键词: 高维稀疏二态数据, 对象组相似度, 对象组特征向量, 聚类, 分类

Abstract: The clustering for high-dimensional sparse data is one of important problem which need to be solved in the field of data mining. Most of the traditional clustering methods do not adapt to high-dimensional data. A clustering method is proposed based on set similarity (SS) and object set feature (OSF), and according to the supremum and infimum of clustering set, the new object can be distributed to different clusters. The idea of this kind method is very clear and easily implemented with reliable and exact results.

Key words: High-dimensional sparse binary data, Object set similarity, Object set feature, Clustering, Classification

中图分类号: 

• TP18