基于属性自表达的低秩超图属性选择算法*

针对高维数据具有低秩形式和属性冗余等特点,提出一种基于属性自表达的无监督超图属性选择算法。具体地,该算法首先利用属性自表达特点用其他属性稀疏地表达每个属性,此自表达形式使用低秩假设寻找高维数据的低秩表示,然后建立超图正则化因子保持高维数据的局部结构,最后利用稀疏正则化因子进行属性选择。属性自表达特性确定属性的重要性,低秩表示相当于考虑数据的全局信息进行子空间学习,超图正则化因子考虑数据的局部结构对数据进行子空间学习。该算法实际上考虑数据全局和局部信息进行子空间学习,更是一种嵌入了子空间学习的属性选择算法。实验结果表明,该算法相比其它对比算法,能更有效地选取属性,并能取得很好的分类效果。

Low rank hypergraph feature selection algorithm based on the self-representation

Due to that high-dimensional data usually is low-rank and contains redundant features, this paper proposed a novel unsupervised hypergraph feature selection algorithm based on self-representation property of features. First, we consider the self-representation matrix to sparsely represent each feature by a linear combination of other features. Such self-representation property is then enforced a low-rank assumption to learn the low-rank representation of high-dimensional data, via considering the global structure of the data to conduct subspace learning. Second, we consider the local structure of the data by a hypergraph based regularizer. In this way, the proposed method integrates subspace learning into the framework of feature selection. Experimental results demonstrated that the proposed can select the best discriminative features and achieve the best classification performance, compared to the competing methods.