PCovR2: A flexible principal covariates regression approach to parsimoniously handle multiple criterion variables

Principal covariates regression (PCovR) allows one to deal with the interpretational and technical problems associated running ordinary using many predictor variables. In PCovR, variables are reduced a limited number of components, simultaneously, criterion regressed on these components. By means weighting parameter, users can flexibly choose how much they want emphasize reconstruction prediction. However, when datasets contain variables, PCovR face new problems, because weights will be obtained some criteria might unrelated predictors. We therefore propose PCovR2, which extends by also reducing few These components predicted based The PCovR2 parameter again used focus predictors criteria, or filtering out relevant predictable compare two other approaches, partial least squares (PLS) principal (PCR), that reduce called PLS2 PCR2. simulated example, we show outperforms PCR2 aims recover all Moreover, conduct simulation study evaluate well succeed in finding (1) underlying (2) subset Finally, illustrate use empirical data.