Weighted Orthogonal Components Regression Analysis

In the linear regression setting, we propose a general framework, termed weighted orthogonal components (WOCR), which encompasses many known methods as special cases, including ridge and principal regression. WOCR makes use of monotonicity inherent in to parameterize weight function. The formulation allows for efficient determination tuning parameters hence is computationally advantageous. Moreover, offers insights deriving new better variants. Specifically, advocate assigning weights based on their correlations with response, may lead enhanced predictive performance. Both simulated studies real data examples are provided assess illustrate advantages proposed methods.