Slice weighted average regression

Abstract It has previously been shown that ordinary least squares can be used to estimate the coefficients of single-index model under only mild conditions. However, estimator is non-robust leading poor estimates for some models. In this paper we propose a new sliced least-squares utilizes ideas from Sliced Inverse Regression. Slices with problematic observations contribute high variability in easily down-weighted robustify procedure. The simple implement and result vast improvements models when compared usual approach. While was initially conceived mind, also show multiple directions obtained, therefore providing another notable advantage using slicing squares. Several simulation studies real data example are included, as well comparisons other recent methods.