Overlapping Sliced Inverse Regression for Dimension Reduction
This is an incremental improvement for researchers in statistical learning and dimension reduction, enhancing the SIR method for better estimation in supervised settings.
The paper tackled the problem of accurately estimating the effective dimension reduction space and number of factors in supervised dimension reduction by proposing an overlapping slicing scheme for sliced inverse regression (SIR), resulting in improved accuracy as verified through simulations and real applications.
Sliced inverse regression (SIR) is a pioneer tool for supervised dimension reduction. It identifies the effective dimension reduction space, the subspace of significant factors with intrinsic lower dimensionality. In this paper, we propose to refine the SIR algorithm through an overlapping slicing scheme. The new algorithm, called overlapping sliced inverse regression (OSIR), is able to estimate the effective dimension reduction space and determine the number of effective factors more accurately. We show that such overlapping procedure has the potential to identify the information contained in the derivatives of the inverse regression curve, which helps to explain the superiority of OSIR. We also prove that OSIR algorithm is $\sqrt n $-consistent and verify its effectiveness by simulations and real applications.