Robust Principal Tensor Component Analysis

As a fundamental and popular tool for data analysis and dimensionality reduction, principal component analysis (PCA) plays an important role in a wide range of disciplines. Due to PCA’s sensitivity to sparse noise, robust PCA formulates a data matrix as the superposition of a low-rank component and a sparse component. When dealing with the ubiquitous multidimensional data, matrix transformation operation is inevitable, which will cause the loss of structure information. Therefore, robust principal tensor component analysis (RPTCA) is proposed, which separates the low-rank and the sparse tensor from the whole tensor by exploring the multidimensional structure properties.