A Survey on High-Dimensional Subspace Clustering

With the rapid development of science and technology, high-dimensional data have been widely used in various fields. Due to complex characteristics data, it is usually distributed union several low-dimensional subspaces. In past decades, subspace clustering (SC) methods studied as they can restore underlying perform fast with help self-expressiveness property. The SC aim construct an affinity matrix by self-representation coefficient then obtain results using spectral method. key how design a model that reveal real structure data. this survey, we focus on two decades present new classification criterion divide them into three categories based purpose clustering, i.e., low-rank sparse SC, local preserving kernel SC. We further subcategories according strategy constructing representation coefficient. addition, applications face recognition, motion segmentation, handwritten digits speech emotion recognition are introduced. Finally, discussed interesting meaningful future research directions.