一种基于聚类集成的无监督特征选择方法

提出了一种无监督的特征选择方法，其基本思想是利用聚类来指导特征选择，对于无类别标签的数据样本集，先进行聚类获得数据类标签，再利用ReliefF算法进行特征选择．采用聚类集成方法解决一些聚类结果的不稳定问题，最终特征选择结果通过多次特征选择综合得到．实验结果表明，该算法具有良好的特征选择性能，在去除无关或冗余特征后可进一步提高聚类质量．

A Clustering Ensemble Based Unsupervised Feature Selection Approach

An unsupervised feature selection approach is proposed,which utilizes clustering to obtain the class label of data object and uses ensemble technique to resolve the instability of clustering.As clustering results generated by some algorithms are usually different from each other,feature selection performs multiply and all results are combined to produce final selected features.In addition,ReliefF is ameliorated,which is a supervised feature selection algorithm and is employed as an essential part in the approach.Experimental results show that the approach can remove redundant features and improve the quality of clustering.